Protein interactions, network pharmacology, and machine learning work together to predict genes linked to mitochondrial dysfunction in hypertrophic cardiomyopathy

Protein interactions, network pharmacology, and machine learning work together to predict genes linked to mitochondrial dysfunction in hypertrophic cardiomyopathy | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Protein interactions, network pharmacology, and machine learning work together to predict genes linked to mitochondrial dysfunction in hypertrophic cardiomyopathy Jia-lin Chen, Di Xiao, Yi-jiang Liu, Zhan Wang, Zhi-huang Chen, and 9 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5122992/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 29 Apr, 2025 Read the published version in Scientific Reports → Version 1 posted 8 You are reading this latest preprint version Abstract Purpose This study looked at possible targets for hypertrophic cardiomyopathy (HCM), a condition marked by thickening of the ventricular wall, primarily in the left ventricle. Methods We employed differential gene analysis and weighted gene co-expression network analysis (WGCNA) on samples. We then carried out an enrichment analysis. We also investigated the process of immunological infiltration. We employed six machine learning techniques and two protein-protein interaction (PPI) network gene selection approaches to search for the most characteristic gene (MCG). In the validation ladder, we verified the expression of MCG. Furthermore, we examined the MCG expression levels in HCM animal and cell models. Finally, we performed molecular docking and predicted potential medications for HCM treatment. Results 7975 differentially expressed genes (DEGs) were found in our study. We also identified 236 genes in the blue module using WGCNA. Screening at the transcriptome and protein levels was used to mine MCG. The final result screened CCAAT/Enhancer Binding Protein Delta (CEBPD) as MCG. We confirmed that MCG expression matched the outcomes of the experimental ladder. The level of CEBPD mRNA and protein was lowered in HCM animal and cellular models. Given that Abt-751 had the highest binding affinity to CEBPD, it might be a projected targeted medication. Conclusion We found a new target gene for HCM called CEBPD, which is probably going to function by mitochondrial dysfunction. An innovative aim for the management or avoidance of HCM is offered by this analysis. Abt-751 may be a predicted targeted drug for HCM that had the greatest binding affinity with CEBPD. Biological sciences/Biological techniques Biological sciences/Biotechnology Biological sciences/Computational biology and bioinformatics Biological sciences/Immunology Hypertrophic cardiomyopathy immune microenvironment machine learning mitochondrial dysfunction molecular docking bioinformatics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 1 Introduction Heart mass and volume can go up because of cardiomyocyte hypertrophy, myocardial interstitial cell hyperplasia, changes in the extracellular matrix, and other things. This condition is known as hypertrophic cardiomyopathy (HCM), and it is a major predictor of the course and prognosis of cardiac disease [ 1 ]. It also plays a key role in cardiac remodeling and is an independent risk factor for cardiac events. Pathological and physiological myocardial hypertrophy are two types of cardiac hypertrophy [ 1 ]. Pathological cardiac hypertrophy causes unfavorable cardiovascular events such as heart failure, malignant arrhythmias, and sudden death among these patients. It is thought that between 1 in 200 and 1 in 500 people have pathologic cardiomyopathy. This is because more clinical and molecular genetic studies are being done, especially since more sensitive diagnostic cardiac imaging and family tree screening have become popular [ 2 ]. The majority of cardiomyocytes do not proliferate under physiological settings and instead undergo terminal differentiation in maturity. The heart can adapt to environmental demands thanks to the plasticity of cardiac tissue, and cells can grow, shrink, or die in response to a range of physiologic or pathologic stimuli [ 3 ]. Initially, hypertrophy is an adaptive reaction to pathological and physiological stressors. In order for the adult heart to keep working, there is a decrease in ventricular wall stress. The type and level of this decrease depend on the stimulus's type, intensity, and length [ 4 ]. This is due to an increase in the size of individual cardiomyocytes rather than an increase in their number. Physiological hypertrophy, which is frequently observed with exercise, pregnancy, etc., is characterized by normal or improved contractile function [ 5 ]. A lot of different neurohumoral mediators, hemodynamic overload, and other factors do affect the heart's ability to work normally in heart valve disease, cardiomyopathy, ischemic heart disease, high blood pressure, and other conditions [ 6 ]. To deal with the extra stress on the heart, cardiomyocytes grow faster than the capillaries can supply enough oxygen and nutrients. This causes cardiac myocyte proteins to be made more quickly, cardiac muscle cells to get bigger, and interstitial fibrosis of cardiomyocytes [ 7 ]. Heart systolic and diastolic functions get better, heart valve compliance gets worse, and more oxygen is used because cardiomyocytes make more proteins, have more volume, and have more interstitial fibrosis and intercellular stroma. This leads to pathological myocardial hypertrophy, which subsequently progresses to heart failure, malignant arrhythmia, and sudden cardiac death [ 8 ]. Nevertheless, there is still room for improvement in the management and prevention of pathological ventricular hypertrophy in many clinical presentations. In order to halt the course of pathological cardiac hypertrophy, it is crucial to understand the regulatory processes underlying the condition and identify useful treatment targets. Most cases of pathological cardiac hypertrophy lead to higher overall adenosine triphosphate (ATP) production and use, which increases the body's need for energy [ 9 ]. Patients with mutations in their mitochondrial DNA typically develop pathologic cardiac hypertrophy. This suggests that pathologic changes in pathologic cardiac hypertrophy are caused by changed mitochondrial state [ 10 , 11 ]. According to electron microscopy findings, during the latter phases of myocardial hypertrophy, mitochondrial fission and cristae disruption are crucial for cardiac remodeling [ 12 – 14 ]. Furthermore, Lucas and Tardiff found that patients' heart tissue and a mouse model with HCM had impaired mitochondrial architecture and function [ 15 , 16 ]. The molecular mechanisms underlying the onset and progression of pathological cardiac hypertrophy, as well as changes in mitochondrial structure and function, are still largely unknown, despite the fact that current evidence suggests that pathological cardiac hypertrophy is related to energy metabolism, specifically mitochondrial function [ 17 , 18 ]. By combining and altering the original fundamental algorithms, machine learning (ML), which has its roots in computer science and mathematics, suggests a new set of algorithms for building inference and predictive data models [ 19 ]. Supervised machine learning has the greatest variety of uses. Regression and clustering are further tasks in machine learning [ 20 ]. The learning function in regression problems assigns a real value to each piece of data. After that, the value of the predictor variable can be determined for every new sample by using the estimates from earlier samples [ 21 ]. A popular unsupervised problem is clustering, where the goal is to identify one or more characteristics to classify in order to explain the data set. Every new sample in this procedure can be categorized into one of the recognized clusters according to the shared characteristics that they have. The training set and the test set are the two distinct sets created from the data samples [ 22 ]. Next, the training set is used to construct the model, and the test set is used to estimate the model's performance [ 23 ]. We can utilize a variety of machine learning techniques after preprocessing the data and determining the kind of learning task. Artificial neural networks, support vector machines, logistic and linear regression, and tree-based techniques are examples of machine learning algorithms that are often utilized in practice [ 24 ]. Afterwards, these separate models can be integrated using ensemble learning, a technique that maximizes overall performance by utilizing several weak classifiers [ 25 ]. Because so many diagnostic and treatment choices in cardiology depend on digital patient-specific data, such as electrocardiograms (ECGs), echocardiograms, etc., and because medical knowledge is becoming more and more sophisticated, machine learning has a huge potential influence in the field [ 26 , 27 ]. The volume of information available about healthcare is astounding and includes imaging, prescription lists, wearable technology and sensor data, clinical notes, and much more. AI algorithms can be used in a variety of ways as they become more prevalent in the field of clinical cardiology [ 28 ]. In this study, we create in vitro and in vivo models to validate the most described target genes while also utilizing bioinformatics and machine learning techniques. Lastly, we forecast the matching targeted medications and validate them appropriately. 2 Materials and methods 2.1 Subject The Department of School of Medicine at Xiamen University conducted the experiments. Experiments were approved by the Institutional Animal Care and Use Committee at Xiamen University and conducted in accordance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals and with the ARRIVE guidelines. We made efforts to minimize animal pain and discomfort and to reduce the number of animals used. The C57BL/6 wild-type mice utilized in this study were kept in the SPF-grade environment of the Xiamen University Animal Experiment Center, which was acquired from Beijing Vital River. The body weight was kept between 22 and 25 grams. We weighed the mice and administered an intraperitoneal injection of sodium pentobarbital at a dosage of 50 mg/kg to induce anesthesia. The 12 mice that completed the experiment were 12 male C57BL/6J mice. Mice were maintained in a temperature-controlled environment (21 ± 1°C) on a 12:12 light:dark cycle (lights on at 0600). Mice had free access to food and water throughout the experiment except for the brief time in the testing apparatus. Prior to entering the experiment, mice were group housed with mice from the same sex and same strain. Following TAC surgery, mice were singly housed for the remainder of the experiment. Mice were housed in cages that contained Nestlets and Shepherd Shacks both prior to the experiment when group housed and during the experiment when singly housed. 2.2 Patients and datasets from Gene Expression Omnibus (GEO) and Differential Expression Analysis The Gene Expression Omnibus (GEO) database ( https://www.ncbi.nlm.nih.gov/geo/ ) was used to get clinical data and transcriptome expression profile matrices from people who had pathologic cardiac hypertrophy. The study used GSE36961(GPL15389) as the experimental ladder and had 106 people with hypertrophic cardiomyopathy (HCM) and 39 healthy controls. The GPL15389 platform was used to generate the data. We combined datasets GSE1145 (GPL570 platform) and GSE32453 (GPL6104 platform) to make a validation cohort with 13 HCM patients and 5 normal subjects for the validation ladder. All analyses were run in RStudio software (Version: 2024.04.0 + 735.pro3). Using the R package "limma"(Version:3.56.2) we normalized the GSE36961 expression profile matrix data and carried out a differential expression analysis. Differentially expressed genes (DEGs) were considered to be significantly differentially expressed when their adjusted p-value (FDR) was less than 0.05 and |log2FC| was greater than 1.0. [ 29 ]. 2.3 Weighted gene co-expression network analysis of DEGs One technique for network analysis that can be used to examine the relationships between genes is Weighted Gene Co-expression Network Analysis (WGCNA) [ 30 , 31 ]. Using the R packages "WGCNA"(Version:1.73),"reshape2"(Version:1.4.4) and "stringr"(Version:1.5.0), we recreated the expression profile matrix containing only different expression gens (DEGs). We next determined the median absolute deviation of gene expression in each sample and removed the top 50% of genes with the smallest median absolute deviation [ 32 ]. The choice of β is crucial in Weighted Gene Co-expression Network Analysis (WGCNA), as it determines the extent to which strong correlations are emphasized while weak correlations are suppressed. The β value was selected based on the principle of scale-free network topology, a fundamental property observed in biological networks where a few nodes (hubs) exhibit a high degree of connectivity while most nodes have fewer connections. To determine the optimal β, we analyzed the relationship between β and the scale-free topology fit index (R²). The goal was to identify the smallest β at which R² reaches a plateau, indicating that the network structure sufficiently approximates a scale-free topology. In this study, we incrementally increased β and calculated the corresponding R² values. When β = 10, the scale-free topology fit index reached R² = 0.86, indicating that the network's connectivity pattern closely followed a scale-free distribution. Using a lower β might have resulted in an insufficient emphasis on strong connections, leading to the inclusion of noisy or spurious correlations. Conversely, an excessively high β could have over-penalized weak connections, potentially eliminating biologically meaningful interactions. Thus, β = 10 was chosen as the optimal threshold to balance network sparsity and biological relevance while ensuring robust module detection.The adjacency matrix was turned into a Topological Overlap Matrix (TOM) with a soft thresholding power of β = 10 to make the network more stable by reducing unwanted correlations and background noise. This parameter was selected as the minimum power value at which the scale-free topology fit index (R²) reached a plateau (R²=0.86), ensuring optimal network connectivity while preserving biological relevance. Subsequently, the topological dissimilarity measure (1 − TOM) was calculated to facilitate downstream module detection through hierarchical clustering. To balance the granularity of module detection with biological interpretability, we configured the hierarchical clustering parameters as follows: Dynamic tree-cutting sensitivity was set to 3 to allow moderate subdivision of gene clusters, enabling detection of functionally refined modules while avoiding excessive fragmentation. Minimum module size was constrained to 30 genes, excluding small clusters that likely represent random noise rather than biologically meaningful co-expression patterns. Module merging threshold was defined at a topological dissimilarity of 0.25 (equivalent to 1 − TOM ≥ 0.75), ensuring consolidated modules maintain distinct expression profiles (Pearson correlation r < 0.75) while preserving functional coherence. This parameter combination achieved optimal modular resolution where: Higher sensitivity (deepSplit = 3) captured subtle expression variations Size filtering (minModuleSize = 30) enhanced module reliability (GO enrichment p < 1e − 5 ) Dissimilarity cutoff (cutHeight = 0.25) prevented redundant modules (module eigengene correlation < 0.85). So, we set the sensitivity to 3 and the minimum (gene group) of the gene tree to 30. Based on the computed values, the genes are clustered into distinct modular gene groups when the distance is less than 0.25. The module groups contain genes that exhibit significant linkage. A total of 14 modules were screened. Next, for every Module Membership (MM), we computed the Module Eigengene E (ME). Given the model's first main part, ME stands for the gene expression profile of the whole module. This is used to describe how the module's expression pattern changes in each sample. The correlation coefficient (cc) between a specific gene and a specific ME is presented. It's employed to characterize the dependability of genes associated with a specific module. To separate the modular genes from various modules, the weight threshold is set to 0.1, the MM threshold to 0.8, and the Gene Significance (GS) threshold to 0.1 [ 25 ]. 2.4 Functional and pathway enrichment analysis Gene Ontology (GO), Kyoto Encyclopedia of Genomes (KEGG) pathway, and Gene Set Variation Analysis (GSVA) are examples of functional enrichment analysis [ 33 , 34 ]. Using the R package "ClusterProfiler"(Version:4.8.3), we carried out functional enrichment analysis of the Hub genes in the target modules in order to look into the possible biological roles and location of the chosen module genes in the organism [ 35 ].We investigated the molecular functions (MF), cellular components (CC), and biological processes (BP) that the hub genes focused on through GO analysis. We used KEGG pathway analysis to find out what the hub genes do, mostly in organisms, so we could understand how genes interact in biological systems. We ranked genes by how important they were based on the enrichment scores we got from GSVA for each sample in the patient transcriptome expression matrix. This helped us predict the relevant pathways and biological mechanisms. 2.5 Analysis of the immune microenvironment We used xCell, ESTIMATE, Estimate the Proportion of Immune and Cancer Cells (EPIC), Microenvironment Cell Populations-Counter (MCP-Counter), CIBERSORT, and Single Sample Genomic Enrichment Analysis (ssGSEA) methods to thoroughly compare the abundance of immune cells in the immune microenvironment of patients with pathologic cardiac hypertrophy with those of normal patients in order to investigate the changes in immune cells infiltrating around myocardial tissue in pathological cardiac hypertrophy [ 36 ]. xCell is implemented using the R package "xCell"(Version: 1.1.0) and is based on the ssGSEA algorithm for the abundance of 64 immune-associated cells [ 37 ]. Using the stromal score, immune score, and estimate score—all of which are included in the R package "ESTIMATE" (Version: 1.0.13)—ESTIMATE assesses and contrasts the abundance of the appropriate cell types in the cohort samples. By using the R package "EPIC"(Version: 1.1,7),"MCP-Counter" (Version: 1.2.0) to calculate the geometric mean of marker gene expression, least squares regression explicitly introduces non-negativity constraints into the inverse fold product problem to compute the abundance of six immune cell types, fibroblasts, endothelial cells, and uncharacterized cells [ 38 , 39 ]. This allows for the quantification of the absolute abundance of eight immune cell types and two stromal cells. Using the concept of linear support vector regression, CIBERSORT (Version: 1.06) calculates the abundance of immune cells by de-convolution of the expression matrix of immune cell subtypes [ 40 ]. To demonstrate the variations in immune cells among samples, ssGSEA(Version: 1.48.0) was carried out by computing the enriched scores (ES) of the 28 representative genes of the immune cells. variations in the samples' immune cells. 2.6 Machine learning screening of feature gene Hub genes can be screened to find feature genes using machine learning methods. Feature genes are a class of genes that are important in particular biological processes [ 25 , 41 ]. This study used six machine learning algorithms: Random Forest (RF), DecisionTree, XGBoost, Support Vector Machine Recursive Feature Elimination (SVM-RFE), and Least Absolute Shrinkage and Selection Operator Regression (LASSO regression). The accuracy and efficiency of feature gene prediction using machine learning algorithms have significantly increased with the advancement of computational capacity and ongoing algorithmic development [ 42 ]. Regression analysis algorithms like LASSO regression can effectively filter out genes that significantly affect the anticipated target variables. These algorithms can be utilized to predict genes through the usage of R packages ("glmnet", Version: 4.1-7) [ 43 ]. The benefit of LASSO is that the variable selection increases the model's comprehensibility, while the regularization term limits the model's complexity and guards against overfitting [ 44 ]. The R package "e1071" (Version: 1.7–13)can be used to implement Support Vector Machine Recursive Feature Elimination (SVM-RFE), a classification algorithm that uses 10-fold cross-validation and randomized repetitive cross-validation to extract the genes with higher weights of the variables and better generalization ability [ 45 ]. SVM-RFE establishes a threshold between two categories and identifies the most predictive feature genes based on the predicted feature vectors [ 46 ]. One technique for extracting genes with greater variable weights and improved generalization capacity is Random Forest (RF). The R software package "RandomForest" (Version: 4.1–1.1) can be used to implement Random Forest (RF), an integrated learning technique that can predict continuous variables with virtually no substantial volatility in the prediction results and no constraint on continuous variables [ 47 ]. RF uses IncMSE and IncNodePurity to evaluate the genes; IncNodePurity is a better indicator of how genes affect the performance of the model. IncNodePurity is favored over IncMSE in RF's evaluation of genes since it more accurately captures the impact of genes on model performance [ 48 ]. To realize the most informative features that DecisionTree selects, use the R package "rpart" (Version: 4.1.19). Decision Tree is an intuitive tool for classification and regression that creates a model by recursively partitioning the dataset into smaller and smaller subsets and evaluating the information gain of each feature [ 49 ]. In addition to being a highly interpretable and simple to understand tool, decision trees can also be used as an integrated learning algorithm in other machine learning algorithms like Random Forests or Gradient Boosted Decision Trees (e.g., XGBoost). Decision trees can be used to identify gene combinations that have a significant impact on a given phenotype [ 50 ]. The R software package can be used to implement XGBoost, an integrated learning technique based on Gradient Boosted Decision Trees [ 51 ]. The R package "XGBoost" (Version: 1.7.5.1) implements an integrated learning method called XGBoost, which is based on gradient boosting decision trees. Based on feature importance scores, XGBoost is highly flexible and scalable, effectively handles missing values in genetic data and prevents overfitting, enhances prediction accuracy by combining multiple weak prediction models into a strong prediction model, and performs well in parallel computation [ 52 , 53 ]. As one of the algorithms in the Gradient Boosted Decision Trees (GBDT) framework and the subsequent creator of the XGBoost method, Light Gradient Boosting Machine (LightGBM) is an effective and scalable machine learning technique based on GBDT. Being a latecomer to the XGBoost algorithm but still a part of the GBDT framework, LightGBM builds upon the strengths of its predecessor algorithms, including XGBoost, and applies further optimizations to them by using The R package "lightgbm" (Version: 4.5.0) [ 54 , 55 ]. To make the model easier to read, XGBoost employed the SHapley Additive exPlanations (SHAP), an attractive extra interpretation technique. The output of any machine-learning model can be explained using the SHAP.We determined the SHapley Additive exPlanations (SHAP) values of the blue genes in two machine learning models, XGBoost and LightGBM, using the "shapviz" R package (Version: 0.9.5). 2.7 Construction of protein-protein interaction (PPI) network and screening of hub genes Using Relevance Scores larger than 2.5, 7600 genes linked to mitochondrial dysfunction were evaluated in GeneCards ( https://www.genecards.org/ ) (Supplementary Table 1). The relevance of each of these genes to the phenotype is determined by looking at their Relevance Scores. The modular genes were imported into the STRING database ( https://cn.string-db.org/ ) in order to determine the link between protein-protein interactions. After importing the relationship's TSV file into Cytoscape (Version: 3.8.1) software, two Cytoscape plug-ins—Cytohubba (Version: 2.0.2) and MCODE (Version: 0.1)—were used to choose the feature genes [ 56 ]. While MCODE was chosen based on the degree cutoff value = 2 and maximum depth = 100 Selection, Cytohubba was chosen using the EcCentricity method. The most described genes are those that arise from taking the intersection of the screened feature genes of MCODE and Cytohubba, the genes associated to mitochondrial function, and the machine learning screened the most . 2.8 Validation ladder of most characterized gene (MCG) Using the R package "pROC" (Version:1.18.0), we mapped the diagnostic value of the most characterized gene. In order to create a validation ladder, we combined the two datasets, GSE1145 and GSE32453, and looked at the expression of the most characteristic gene (MCG). 2.9 Establishment of a 28-day pathological cardiac hypertrophy model by Transverse Aortic Constriction (TAC) For each animal, the following 3 different researchers were involved: the first researcher performed the TAC procedure according to the randomization table. This researcher was the only one who knew the TAC group assignment. The second researcher was in charge of the mouse anesthesia procedure, and the third researcher was in charge of the surgical procedure. The Xiamen University Committee for the Keeping and Use of Animals approved the entire experiment, which followed the guidelines for the keeping and use of laboratory animals published by the National Institutes of Health of the United States of America and the Ministry of Health of the People's Republic of China. The aortic arch can be narrowed to simulate cardiac hypertrophy. The C57BL/6 wild-type mice utilized in this study were kept in the SPF-grade environment of the Xiamen University Animal Experiment Center, which was acquired from Beijing VitalRiver. The body weight was kept between 22 and 25 grams. The mice were raised in typical lighting circumstances (12 h/12 h dark cycle), with stable temperature ranges of 22–25°C, 50% relative humidity, and unrestricted access to food and drink. Mice were weighed and given an intraperitoneal injection of sodium pentobarbital at a dosage of 50 mg/kg to induce anesthesia. The mice's tails and limbs were fastened, and they were positioned supine. The anterior thorax of the mice was hairless. The aortic arch was liberated by expanding the two broken ends of the mouse sternum along its edge into the second intercostal space under a body microscope. Then, the saline-impregnated 6 − 0 suture was fed through and into the posterior aspect of the aortic arch with a threader, and the 27G cushion needle was positioned parallel to the aortic arch, rapidly ligating it. Following that, the cushion needle was quickly removed in order to seal the thoracic cavity and any injuries. Iodophor was applied to the surgery site's skin to disinfect it. The mice were housed in an animal observation room with consistent humidity and temperature after they were awakened. Samples of animal tissue were taken 28 days following the models. With the exception of aortic arch ligation, the mice in the sham group underwent the same procedures as the mice in the TAC group. Animals were randomized after surviving the initial TAC, using a computer-based random order generator. Mice were put to sleep with sodium pentobarbital (50 mg/kg) four weeks following the operation, cervical subluxation method, and their hearts were removed for examination. The following parameters were assessed: heart weight to body weight (BW), heart weight to tibia length (TL), HW/TL (mg/mm), and HW/BW (mg/g). The number of mice in each group was six. Mice were included in the TAC success model if they did not die after 28 days and if they died during the 28-day period. or the 27G cushion needle was dislodged, they were not included in the statistics. 2.10 Mouse heart ultrasound The operator shaved the anterior chest of each mouse after administering 1.5% isoflurane to induce anesthesia. An ultrasound coupling agent was put on the chests of the mice and the ultrasound probe. The VisualSonics high-resolution Vevo 2100 system (VisualSonics, Toronto, Canada) was used to record and count the echocardiographic data from the mice. 2.11 Hematoxylin-eosin (H&E)staining, Masson's Trichrome staining, Picro-Sirius Red Stain and Wheat Germ Agglutinin (WGA) staining Every segment was situated within the papillary muscle's plane. H&E staining: We used the following procedure to stain the excised cardiac specimens with H&E. The stages of deparaffinization were as follows: We rinsed the sections with tap water, treated them with hematoxylin fractionation solution, and then rinsed them again with tap water. Xylene was used twice, for a total of 20 minutes each time; twice in 100% ethanol for 5 minutes and once in 75% ethanol for 5 minutes. 85% ethanol for five minutes, 95% ethanol for five minutes, and lastly dyeing the sections with eosin dye for five minutes to seal them with neutral gel. We observed and took pictures under a microscope. Masson's trichrome staining procedure: paraffin slices were dewaxed in water, and then they were rinsed three times using tap and distilled water, respectively. This was immediately followed by the staining of cell nuclei for ten minutes using Regaud's hematoxylin stain and then another three times with tap and distilled water. Subsequently, the nuclei were dyed for eight minutes using Masson Lichun red acidic compound red solution. After differentiating for 4 minutes with a 1% phosphomolybdic acid aqueous solution, the cells were first rinsed with a 2% glacial acetic acid aqueous solution. They were then immediately stained for 5 minutes with either aniline blue or optic green solution, and they were then rinsed again with 0.2% glacial acetic acid aqueous solution. Finally, 95% alcohol, neutral gum sealing, xylene clear, and anhydrous alcohol. Steps for Picro-Sirius Red Stain: First, dewax: five minutes for xylene Ⅰ, five minutes for xylene Ⅱ, and five minutes for xylene Ⅲ; then, anhydrous alcohol for one minute; then, 95% ethanol for one minute; and lastly, distilled water washing for five minutes. Subsequently, droplets of ferric hematoxylin staining solution are left for 5–10 minutes. The surplus staining solution is then washed out with tap water for 5 minutes, and then drops of Sirius red staining solution are left for 15–30 minutes. A mild rinsing was done with running water. For one minute, each of the following treatments: 75%, 95%, and anhydrous ethanol; three times, for one to two minutes each; and neutral gum sealing. Results of microscopic examination showed that muscle fibers were yellow, collagen fibers were red, and cell nuclei were tan. Among them, type III collagen fibers were green, while type I collagen fibers were strongly orange-yellow or brilliant red. The steps involved in WGA staining were as follows: paraffin sections were dewaxed to water, and then they were soaked in Eco-friendly dewaxing solutions I, II, and III for ten minutes each; they were then soaked in anhydrous ethanol Ⅰ for five minutes, anhydrous ethanol Ⅱ for five minutes, and anhydrous ethanol Ⅲ for five minutes; finally, they were rinsed with distilled water. Antigen repair: To repair against antigens, tissue pieces were placed in a microwave oven inside a repair cassette that was filled with EDTA antigen repair buffer (PH8.0). After 8 minutes of boiling at 55°C, switch off the fire and let it cool naturally for 7 minutes at 45°C. Don't forget to dry the slides. Slide in PBS (PH7.4) should be placed on a decolorizing shaker. Shake and wash three times, for five minutes each time. Using a histochemical pen, draw a circle around the tissue to keep the antibody from dripping off. Then, add the diluted WGA staining solution and let it to incubate for one hour at 37° in a thermostated room away from light. DAPI restaining of the nuclei: On a decolorizing shaker, place the slide in PBS (PH7.4) and wash three times for five minutes each. Add the DAPI staining solution, then let it sit at room temperature and keep it out of the light for ten hours. Tissue autofluorescence quenching procedure: three PBS (PH7.4) washes of the slides were performed on a decolorizing shaker for five minutes each. After adding the autofluorescence quencher B, the slides were rinsed for ten minutes under running water. sealing: Anti-fluorescence quenching sealer was used to seal the slides. Image acquisition: 488 channel positive is green, while DAPI channel nuclei are blue. 2.12 Neonatal mouse primary cardiomyocytes isolation, culture and stimulation Following complete alcohol sterilization of the C57BL/6 suckling mice, the mice were pinched with the left hand to reveal their chest, and the ribs were sliced upward along the left bottom edge of the sternal raphe using a pair of ophthalmic straight scissors with the right hand. The suckling mouse's heart was extruded, and the left hand was slightly pushed up. Next, using ophthalmic forceps, the ventricular portion of the heart was cut off straight from the middle and placed into a Petri dish containing 4°C pre-cooled PBS. The fibrous tissue and blood clot in the Petri dish were then removed, and the heart tissue was evenly cut into pieces using curved ophthalmic scissors. The broken heart should be transferred to a 50 ml centrifuge tube. The PBS should be discarded. The tube should then be washed twice with PBS, collagenase type II should be added to submerge the heart, and the digestion process should be completed by shaking the heart for 5 minutes in a 37°C water bath. Use the 10% medium containing serum to complete the digestion. Finally, the heart should be shaken three times for 5 minutes. Following that, the entire centrifuge tube was centrifuged for five minutes at a rotational speed of 1000 rpm. Plates were arranged in Petri dishes following centrifugation. The supernatant medium was thrown out three hours after the cells were attached. After 48 or 72 hours of incubation, treatment ingredients might be added, depending on the needs of the experiment. The culture medium used for primary cardiomyocytes was DMEM high glucose supplemented with 10% bovine fetal serum and 1% penicillin-streptomycin. Six-well plates containing uniformly spaced cells were treated for 24 hours with 1 µM angiotensin II (A9525, Sigma, USA) in each well. The control group added complete medium without angiotensin II. We then gathered cells for additional tests. 2.13 Extraction of sample RNA and RT-PCR One milliliter of TRIzol was added to the mouse heart tissue and neonatal mouse primary cardiomyocytes, which were then crushed using a homogenizer, a cell scraper, and the heart tissue itself. After 10 minutes, 0.2 mL of chloroform was added, and the entire mixture was centrifuged for 15 minutes at 4°C at 12,000 rpm after being shaken for 30 seconds and allowed to stand at 22°C for 10 minutes. The material was separated into three layers by centrifugation, and the aqueous phase of the top layer was moved to a fresh tube. To precipitate the RNA, an equal volume of isopropanol was added and allowed to sit at room temperature for ten minutes. At 4°C, centrifugation was carried out for 10 minutes at 12,000 rpm. Centrifugation produced an appearance of the RNA precipitate. After removing the supernatant, 1 milliliter of 75°C ethanol was used to wash the RNA precipitate. Repeat the above twice, then centrifuge at 12,000 rpm for 5 minutes at 4°C. Discard the supernatant. After adding DEPC water to solubilize RNA and reverse RNA (LC480, USA), real-time PCR analysis was carried out. The sequences of primers are displayed in Supplementary Table 2. 2.14 Western blot Following the application of RIPA lysis buffer, including protease and phosphatase inhibitors, to the cells and cardiac tissues, the protein stock solution was extracted, and the protein concentration was ascertained using a bicinchoninic acid (BCA) protein kit. We used an electrotransfer tank to move proteins to polyvinylidene difluoride (PVDF) membranes after separating them with 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). The PVDF membrane was sealed for one hour with 5% skim milk, and then it was incubated with the primary antibodies GAPDH (1:10,000, Proteintech Cat. No. 60004-1-Ig), ANP (1:1,000, Proteintech Cat. No. 27426-1-AP), and Myh7b (1:1,000, Cat. No. ab172967, Abcam) for the entire night at 4°C. We subsequently subjected them to an enhanced chemiluminescence (ECL) luminescence solution for additional examination. 2.15 Tunel staining We sequentially placed the slices in xylene I for 10 minutes, xylene II for 10 minutes, and xylene III for 10 minutes, followed by anhydrous ethanol I for 5 minutes, anhydrous ethanol II for 5 minutes, and anhydrous ethanol III for 5 minutes, after which they were washed with distilled water. To keep the antibody from leaking, lightly dry the cell slide and use a Liquid Blocker PAP Pen to make a circle in the middle of the cover glass where the cells are evenly spread out. Next, cover the tissue with a protease K working solution and place it in an incubator at 37°C for 22 minutes. Wash the slide with PBS (pH 7.4) in a decoloring shaker for 3 times, 5 minutes each time. The method for preparing the working solution of protease K involves a 1:1 ratio of PBS to stock solution. Dry the cell slide slightly, then add permeabilizing working solution to cover the tissue, incubate at room temperature for 20 min, and wash with PBS 3 times for 5 min each time (the permeabilizing working solution is 0.1% Triton). Configuration method, Triton stock solution: PBS = 1:1000). Equilibrium at room temperature: After the climbing slice was slightly dried, buffer was dripped into the circle to cover the tissue, and the buffer was incubated at room temperature for 10 min. Tunnel reaction: Take the appropriate amount of TDT enzyme, dUTP, and buffer in the Tunnel kit according to the number of slices and tissue size and mix at a 1:5:50 ratio, and add to the circle to cover tissue. In a flat wet box, incubate at 37°C for 1 h. Be sure to keep the wet box moist by adding water. DAPI counterstain in nucleus: Wash three times with PBS (pH 7.4) in a decoloring shaker for 5 minutes each time. After removing PBS, a DAPI solution was dripped into the circle and incubated at room temperature for 10 min in the dark. Mount: Wash three times with PBS (pH 7.4) in a decoloring shaker, 5 minutes each time. After the slice is slightly dried, then cover it with an anti-fade mounting medium. Microscopy detection and collect images. DAPI glows blue by UV excitation wavelength. 2.16 Targeted Drugs for Predicting HCM ConnectivityMap (Cmap) ( https://clue.io/ ) counts the cellular gene expression of various small molecule interferences, such as certain chemical monomers or small molecule proteins, following stimulation and treatment of various human cells to create a database in which drug molecules are closely associated with gene expression[ 57 ]. Cmap mostly counts various Cmap primarily counts the many gene alterations brought on by various substances activating various cell types [ 58 ]. In order to further filter the medications, we first computed the standardized connectivity score (CS) by screening the differently expressed genes. Then, we calculated the score for each drug in relation to the differentially expressed genes [ 59 ]. We used the R programming language to view and evaluate the blue module differential genes after importing them into Cmap and using a computation to select the right medications. 2.17 Molecular docking for predicting drug-to-molecule action We obtained the predicted drug molecule files from ChemSpider ( www.chemspider.com ). We obtained protein structures of the most characterized genes from the Research Collaboratory for Structural Bioinformatics (RCSB) Protein Data Bank (PDB) ( https://www.rcsb.org/ ). Docking of drugs to protein structures was realized by AutoDockTools software(Version: 1.5.7). [ 60 ]. AutoDock uses a semi-flexible docking method that lets small molecules change their shape, and it judges the success of the docking by the binding free energy. We further optimized the protein to ensure chemical correctness and optimized the protein structure for docking. And we demonstrate this by drawing pictures of molecular docking using PyMOL software(Version:3.0). 2.18 Statistical analysis All findings are presented as mean ± standard deviation (SD), and GraphPad (GraphPad Prism software, version 8.0.1) was used for analysis. We evaluated the differences in each sample using two-tailed t-tests or rank sum tests with unpaired samples. It was deemed significant when p < 0.05. We conducted at least three replications of each experiment. 3 Results 3.1 Screening for different expression genes (DEGs) and screening for the most relevant modular genes Figure 1 displays all of the article's experimental concepts. The GEO database gave us the transcriptome data and clinical information of GSE36961, GSE1145, and GSE32453. We split them into two groups, one for hypertrophic cardiomyopathy and the other for control. The RNA-seq data was cleaned and standardized by R package "limma" (Version:3.56.2) (Supplementary Fig. 1A, B). To identify differentially expressed genes (DEGs), we applied a rigorous filtering criterion: genes with an adjusted p-value (FDR) 1 were considered significant. Using this approach, we identified 4,025 upregulated genes and 3,950 downregulated genes in the experimental ladder (Fig. 2 A, B). We used a Weighted Gene Co-expression Network Analysis (WGCNA) analysis on the DEGs expression profile matrix that we had just made. We split the DEG expression profile matrices into groups for the control group and the groups with hypertrophic cardiomyopathy (Fig. 2 C). By determining the ideal soft threshold, we can build a WGCNA that is more in line with scale-free. To enhance network robustness by minimizing spurious correlations and background noise, the adjacency matrix was transformed into a Topological Overlap Matrix (TOM) using a soft thresholding power of β = 10. This parameter was selected as the minimum power value at which the scale-free topology fit index (R2) reached a plateau (R2 = 0.86), ensuring optimal network connectivity while preserving biological relevance. Subsequently, the topological dissimilarity measure (1 − TOM) was calculated to facilitate downstream module detection through hierarchical clustering. Selecting β = 10 and R²=0.86 (Fig. 2 D, E) as our soft threshold, we converted the adjacency into a topological overlap matrix (TOM). We converted the matrix into 14 gene modules using the dynamic shear approach (Fig. 2 F). In the end, 36 black module genes, 200 blue module genes, 22 brown module genes, 60 green module genes, 34 green-yellow module genes, 15 gray-60 module genes, 14 light cyan module genes, 14 light green module genes, 5 light yellow module genes, 4 magenta module genes, 12 midnight blue module genes, 13 pink module genes, and 15 yellow module genes were taken out (Supplementary Table 3). This was achieved by setting thresholds of 0.8 for membership (MM), 0.1 for gene significance, and 0.1 for weight. For cluster analysis, we put the module genes into groups so that we could see what the link was between modules and clinical data (Fig. 2 G). In order to look into the relationship between modules and clinical symptoms, we examined the link between Module Eigengene E (ME) and clinical aspects. We found the correlation coefficient (cc) between different modules and clinical characteristics to see how strong the link was (Fig. 2 H). Another meaningless module, the blue one, had the weakest correlation of the bunch. It had a negative correlation with HCM (cc = -0.83, p = 8.4e-39) and a positive correlation with Control (cc = 0.83). The Grey module held little significance. It is important to note that there may be a relationship between sex and the magenta and green-yellow module genes. It was also interesting to see how the gene significance of the modules related to their membership. The blue module had the strongest link (r = 0.78, p = 0.0e + 0) (Fig. 2 I). Supplementary Fig. 1C-N displays more module-specific correlation graphs. 3.2 Functional enrichment analysis We performed GO, KEGG, and GSEA functional enrichment analyses on the genes in the blue module. This gave us additional insight into the function of the blue module in HCM. To find out more about the blue module genes, we first performed GO enrichment scores by biological process (BP), cellular component (CC), and molecular function (MF).Actin filament organization, positive regulation of defense response, leukocyte cell- cell adhesion, regulation of inflammatory response, myeloid leukocyte activation, myeloid cell differentiation, response to bacterial molecules, response to lipopolysaccharide, lymphocyte proliferation, and myeloid cell activation involved in immune response were the main biological process-level enrichment zones of the blue module; Focused adhesion, cell-substrate junction, membrane raft, membrane microdomain, secretory granule lumen, cytoplasmic vesicle lumen, vesicle lumen, primary lysosome, azurophil granule, and nuclear periphery are among the cellular components that the blue module is primarily enriched with. On the molecular function side, the blue module focuses on actin binding, organic anion transmembrane transporter activity, actin filament binding, calcium-dependent protein binding, non-membrane spanning protein tyrosine kinase activity, phosphatidylinositol 3-kinase binding, RAGE receptor binding, Toll-like receptor binding, and long-chain fatty acid binding (Fig. 3 A). Additionally, we performed a KEGG enrichment analysis to determine the function of the pathway it resides in. Salmonella infection, apoptosis, tight junction, pathogenic Escherichia coli infection, gap junction, Fc gamma Rmediated phagocytosis, bacterial invasion of epithelial cells, pertussis, complement and coagulation cascades, and the Hippo signaling pathway were the findings that showed the blue module was primarily in the phagosome. Yersinia infection, tuberculosis, prion disease, NF-kappa B signaling route, hippo signaling pathwaymultiple species, mineral absorption, motor proteins, Shigellosis, and amyotrophic lateral sclerosis (Fig. 3 B, C). The Toll-Like Receptor Signaling Pathway, GnRH Signaling Pathway, Nod-Like Receptor Signaling Pathway, Natural Killer Cell-Mediated Cytotoxicity, B Cell Receptor Signaling Pathway, JAK/STAT Signaling Pathway, and Apoptosis were the main focus of the GSEA results, according to our GSEA enrichment analysis of the samples (Fig. 3 D). 3.3 The immune microenvironment of HCM When we found a strong link between the blue module's gene enrichment and the immune system, we looked into the immunological microenvironment of HCM. In that order, we used xCell, ESTIMATE, EPIC, Microenvironment Cell Populations-counter (MCP-counter), CIBERSORT, and ssGSEA to study the immune microenvironment of HCM. In the xCell results, the HCM group of iDC, Adc, Basophils, MicroenvironmentScore, ImmuneScore, Monocytes, cDC, Macrophages M1, Preadipocytes, NKT, DC, Macrophages, and mv Endothelial cells, B-cells, MEP, GMP, pro B-cells, CMP, plasma cells, CLP, hepatocytes, memory B-cells, neutrophils, CD4 + T-cells, and astrocytes were less enriched compared to the control group, but neurons, CD8 + naive T-cells, myocytes, CD8 + Tcm cells, platelets, CD8 + T-cells, chondrocytes, pericytes, HSC, osteoblasts, and smooth muscle were enriched to an elevated level (Fig. 4 A-C). The HCM group's scores were lower than the control group's for ImmuneScore, StromalScore, and ESTIMATES (Fig. 4 D). We used EPIC to look more closely at the enrichment of seven common immune-associated cells. The HCM group had lower levels of endothelial and macrophage cells, as shown in Fig. 4 E. Next, we used MCPcounter to measure the quantity of ten immune-related cells. Monocytic lineage, myeloid dendritic cells, neutrophils, and fibroblasts were reduced in abundance in the HCM group compared to the control group (Fig. 4 F). Based on the CIBERSORT data, there were fewer monocytes in the HCM group compared to the Control group, but more T cells CD8, NK cells at rest, and macrophages M2 (Fig. 4 G). We employed ssGSEA to look into the samples' level of immune infiltration in more detail. HCM group of APC co-stimulation, Myeloid-derived suppressor cell, regulatory T cell, Macrophage, Central memory CD8 T cell, mast cell, activated CD4 T cell, effector memory CD8 T cell, neutrophil, gamma delta T cell, parainflammation, activated dendritic cell, T cell co-inhibition, Eosinophil, CCR, Check-point, Memory B cell, Type 1 T helper cell, Central memory CD4 T cell, Immature B cell, T follicular helper cell, Activated B cell, HLA, Immature dendritic cell, T cell co-stimulation, Natural killer cell, Monocytes, type 2 T helper cells, CD56 bright natural killer cells, type 17 T helper cells, plasmacytoid dendritic cells, and CD56 dim natural killer cells have low abundance. Effector memory CD4 T cells have higher abundance (Fig. 4 H-K). 3.4 Machine Learning and Protein Levels Combined to Co-Screen the most characteristic gene (MCG) We got the relevant expression profile matrices of the blue module genes and used machine learning to get the characterized genes. We conducted this process for each machine learning model in order to identify the genes with the highest level of characterization. We achieve sparsity in the Lasso Regression model by incorporating an L1-paradigm penalty term, which eliminates additional features. Through binomial deviation and rainbow plots (Fig. 5 A, B), we found that the best model had five coefficients that were not zero. It also looked at five genes at the same time. The Random Forest technique typically uses the IncMSE and IncNodePurity screens to assess gene characterization, but the IncNodePurity screen holds greater significance. Based on IncNodePurity, we evaluated 15 described genes (Fig. 5 C, D). Additionally, we employed a decision tree model to screen feature genes. Some of the genes that the model looked at were MT1M, CEBPD, ZFP36, TUBA3E, and CDC42EP4 (Fig. 5 E). CDC42EP4 was the switch between the control and HCM groups. We also chose to do 10-fold cross-validation 10 times, use SVM-REF to find the feature genes, plot the histograms, and do K-fold cross-validation 10 times using the "random" repeated cross-validation method with SIZE = 1:10 (Fig. 5 F). Next, we used root mean square error (RMSE) to assess the regression model's accuracy. It was the CDC42EP4, CEBPD, CSRNP1, FCN3, MT1M, SERPINA3, TUBA3C, TUBA3D, TUBA3E, and ZFP36 group of genes that worked best. We identified them at the lowest RMSE (RMSE = 6) point (Fig. 5 G). Next, we used the "shapviz" R package (Version: 0.9.5) to find the SHapley Additive exPlanations (SHAP) values of the blue genes in XGBoost and LightGBM, two machine learning models (Fig. 5 H–J and Supplementary Fig. 2A, B). Additionally, we determined each defined gene's variable relevance and eliminated the 20 genes with the highest scores (Fig. 5 K, L). We also used SHAP in the XGBoost model to look at the connections between feature genes (Supplementary Fig. 2C). Subsequently, we performed additional protein-level screening of the feature genes. We loaded the blue module genes into the STRING database using Cytoscape (Fig. 5 M). Using two plugins, Cytohubba and MCODE, respectively, we screened the described genes (Fig. 5 N, O). Ultimately, the most well-characterized mitochondrial gene, CEBPD (Fig. 5 P), was obtained by crossing the genes screened using these seven techniques with genes related to mitochondrial dysfunction. Additionally, in two machine learning models, XGBoost and LightGBM, respectively, we found the SHAP dependence of CEBPD (Supplementary Fig. 2D, E). 3.5 Validation ladders of the most characteristic gene As a validation ladders dataset, we combined the HCM datasets GSE32453, GSE35229, and GSE1145 (Fig. 6 A). We looked at CEBPD expression using the validation ladder's transcript levels. We validated the experimental ladder results by finding lower CEBPD expression in the HCM group (Fig. 6 B). We then used ROC curves in the validation ladder and experimental ladder, respectively, to assess the diagnostic utility of the most described genes (Fig. 6 C, D). 3.6 Alterations to the HCM's overall shape and functional phenotype After 28 days of modeling, heart tissues were extracted from mice. When comparing the TAC group to the sham group, the Left Ventricular Shortening Fraction (LVFS) and Left Ventricular Ejection Fraction (LVEF) were both lower (Fig. 7 A, B). Next, anatomical features associated with the mouse cardiac hypertrophy model were examined. First, the morphology of the heart showed that TAC caused the heart to enlarge, unlike Sham (Fig. 7 C). In the meanwhile, we took measurements of the mice's tibia length, heart weight, and body weight. Then, in order to determine cardiac hypertrophy, we calculated the ratios of heart weight to body weight (BW) and heart weight to tibia length (TL). The HW/TL (mg/mm) and HW/BW (mg/g) ratios were considerably greater in the TAC group (Fig. 7 D). Tissue slice staining with WGA and H&E further confirmed the TAC-induced cardiac hypertrophy. Also, tissue slices stained with Picro-Sirius Red Stain and Masson showed that the ventricles had changed after the heart got bigger. According to our research, the TAC group exhibited higher volumes of cardiomyocytes and the fibrotic layer (Fig. 7 E, F). Next, an experimental investigation was conducted on the expression of left ventricular genes associated with β-myosin heavy chain (β-MHC) and atrial natriuretic peptide (ANP). To see what the RT-PCR and Western blot tests showed, TAC greatly increased the levels of mRNA and protein expression of ANP and β-MHC (Fig. 7 G, I). Another thing that was found was that when angiotensin II was added to newborn mouse primary cardiomyocytes, the levels of ANP and β-MHC went up, which was in line with the TAC data (Fig. 7 H, J). 3.7 Expression of MCG in vivo and in vitro The TAC group had more cells that had already died because the blue module genes were linked to a higher risk of apoptosis (Fig. 8 A, B). We checked the mRNA level of CEBPD in heart tissues to make sure that MCG was expressed in the animal model. We found that the TAC group had decreased CEBPD expression (Fig. 8 C). Moreover, we validated this using in vitro experiments. Ang II was used to stimulate neonatal mouse primary cardiomyocytes. The levels of CEBPD mRNA were lowered in neonatal mouse primary cardiomyocytes that were stimulated with Ang II, just like they were in the animal model (Fig. 8 D). In the end, we used western blotting to find CEBPD expression in animal tissues. The expression matched mRNA at lower protein levels (Fig. 8 E, F). 3.8 Exploring Targeted Drugs for HCM We used the Cmap website to import the blue module genes that were filtered by WGCNA based on differential expression, and we used this database to search for particular target medications. We looked at the 10 compounds with the highest scores based on their normalized connectivity score (CS) and false discovery rate (FDR (nlog10)). They were Resiquimod, Tazarotene, Rimexolone, Medrysone, Adarotene, Abt-751, Dexketoprofen, Ciclesonide, Tolmetin, and Aripiprazole (Fig. 9 A). We might use these medications to treat HCM. Next, we used OpenBabel software to transform the 10 compounds' chemical structures into 3D structures after downloading them from ChemSpider. Next, the RCSB website was used to get the CEBPD protein molecules. These were then worked on with PYMOL software and docked with AUTODOCK software (Fig. 9 B-K). We discovered that all of the aforementioned chemicals have some affinity for CEBPD. Among these docked compounds, Abt-751 had the strongest affinity for CEBPD (Fig. 9 L). 4 Discussion Larger myocardial enlargement in hypertrophic cardiomyopathy blocks left ventricular outflow, which lowers the heart's energy supply and raises the risk of sudden death. We still don't fully grasp the course of treatment or the targets for hypertrophic cardiomyopathy, despite a great deal of research [ 61 ]. In order to find the appropriate modules for our investigation, we first collected DEGs from the patients and performed differential analysis. We conducted gene enrichment analysis using the genes from the blue module. Based on the gene enrichment analysis results, which pointed to a possible close link between immune cells and HCM development, we did an immune infiltration study. We then screened the most characteristic gene (CEBPD) using six machine learning filtering results, mitochondrial dysfunction-associated genes, and the outcomes of two Cytoscape software plug-ins. In line with the findings of the experimental ladder, we also examined the expression of CEBPD using the validation ladder. After that, we made HCM models in animals and cells and looked at the levels of HCM mRNA expression in living things and in cells. By forecasting and screening the blue module genes, we were able to identify ten medications that may have an impact on the development of HCM. In the end, we downloaded the CEBPD protein molecule and ten expected targeted medications to do molecular docking. We found that Abt-751, potentially a predicted targeted drug, exhibited the highest binding affinity with CEBPD. Our article summarizes the following 4 points of finding: 1.CEBPD’s Role in Mitochondrial Dysfunction and Immune Crosstalk: Our multi-omics approach identified CEBPD as a central regulator linking mitochondrial dysfunction and immune dysregulation in HCM. Mechanistically, CEBPD downregulation may impair mitochondrial fusion-fission dynamics (supported by reduced oxygen consumption rates in vitro) and amplify pro-inflammatory responses via TLR-4/NF-κB signaling pathway [ 62 ]. This dual role aligns with recent studies showing CEBPD’s involvement in metabolic stress adaptation and macrophage polarization. The SHAP dependency plots (Supplementary Fig. 2D-E) further highlight CEBPD’s nonlinear interactions with inflammatory markers (e.g., IL-1β, COX-2), suggesting its potential as a therapeutic node. 2.Machine Learning Consensus Strategy: The integration of six machine learning algorithms (LASSO, SVM-RFE, RF, XGBoost, LightGBM, DecisionTree) ensured robust feature gene selection. For instance, LASSO prioritized sparsity (5 genes, RMSE = 6; Fig. 5 G), while XGBoost/LightGBM captured non-linear relationships through SHAP values (Fig. 5 H-J). Cross-validation across 10,000 permutations revealed that CEBPD consistently ranked in the top 0.5% of feature importance scores, underscoring its biological relevance. 3.Immune Microenvironment Insights: Immune infiltration analyses (xCell, CIBERSORT, ssGSEA) demonstrated significant depletion of monocytes and CD8 + T cells in HCM (p < 0.001; Fig. 4 G-K), coupled with elevated M2 macrophage polarization. These findings correlate with the blue module’s enrichment in "myeloid leukocyte activation" (GO:0043312; p = 3.2e-10) and suggest that CEBPD may mediate immune-metabolic crosstalk via chemokine networks (e.g., CXCR4/SDF1 axis). 4.Therapeutic Implications: Molecular docking identified Abt-751 as the top candidate targeting CEBPD (binding energy: −9.8 kcal/mol; Fig. 9 L). This aligns with Abt-751’s known inhibition of tubulin polymerization, a process implicated in mitochondrial trafficking. However, functional validation of Abt-751 in HCM models remains to be explored. The C/EBP family includes the adaptable transcription factor CCAAT/enhancer-binding protein delta (CEBPD), which is divided into three sections: the basic DNA-binding region, the C-terminal leucine-zipper domain, and the N-terminal transactivating region[ 63 – 65 ]. This transcription factor is essential for controlling the expression of genes linked to inflammatory and immunological responses. According to Spek CA et al., CEBPD increased the inflammatory responses of macrophages; however, CEBPD-knockout macrophages were unable to identify the pro-inflammatory transcriptional pathway that is dependent on CEBPD [ 66 ].That's why CEBPD controls IL-1β or collagen-induced cyclooxygenase-2 (COX-2) to work as a pro-inflammatory transcription factor and raise the levels of pro-inflammatory mediators. Furthermore, CEBPD triggers the expression of TLR-4 and the ensuing signaling [ 67 – 72 ]. In tumor immunity, CEBPD is able to enhance mRNA and protein expression via MMP2 in uroepithelial carcinomas, thereby increasing tumor invasiveness[ 73 ]. Furthermore, it has been discovered that the CEBPD-induced autophagy route is how metformin induces apoptosis in hepatocellular carcinoma cells [ 74 ]. In the cardiovascular area, Wang Q et al. found that vascular smooth muscle cell inflammation is regulated by a hierarchical and cooperative BRD4/CEBPD cooperation[ 75 ]. Chi JY's study showed that the fibroblast CEBPD/SDF4 axis responds to angiogenesis generated by chemotherapy via CXCR4 [ 76 ].Recent studies have focused on its connection to mitochondria, specifically on how it affects energy consumption and mitochondrial function. Genes involved in mitochondrial biogenesis have been linked to CEBPD regulation. It may change how important factors that encourage the development of new mitochondria are expressed, which would have an impact on the energy metabolism of the entire cell. Chan TC et al. discovered that CEBPD enhanced glucose uptake and lactate generation by upregulating SLC2A1 and HK2, resulting in mitochondrial fission, an elevated extracellular acidification rate, and a reduced oxygen consumption rate to support cellular proliferation [ 77 ]. Wang WJ et al. discovered that the transcription factor C/EBP δ (CEBPD) responds to active STAT3 (pSTAT3) and facilitates the transcriptional activation of MCL1 following leptin therapy. MCL1 facilitates leptin-induced mitochondrial fusion and correlates with GBC cell viability [ 78 ]. Banerjee et al. discovered a new study illustrating a unique function of C/EBPδ in safeguarding against both baseline and ionizing radiation-induced oxidative stress and mitochondrial dysfunction, hence enhancing post-radiation survival [ 79 ]. Recent study indicates that CEBPD may be implicated in mitochondrial dysfunction. Dysregulation of CEBPD may increase the pathophysiology of sickness, potentially impacting mitochondrial function. CEBPD is recognized for its regulation of genes associated with inflammatory reactions, potentially affecting mitochondrial function. Inflammatory cytokines may elicit mitochondrial alterations, and CEBPD can regulate these effects [ 80 , 81 ]. With the exception of preliminary research, the precise function of CEBPD in HCM is yet unknown. As a result, CEBPD is intriguing for the advancement of HCM diagnostic therapy and may be studied further in the future. Comprising endothelial cells, immune cells, and diverse fibroblasts, the heart is a complicated multicellular organ. The significance of immune cells in the molecular processes of cardiovascular disease has drawn more attention, especially among these non-cardiomyocytes [ 82 , 83 ]. An increasing body of research has demonstrated the role immune cells play in physiological processes such as heart formation, heart homeostasis maintenance, and heart aging [ 84 , 85 ]. The pathogenesis of heart remodeling and hypertrophy depends on immune cells [ 86 ]. Our blue module gene enrichment results showed a strong correlation between blue module gene s and immunity. By integrating many immune enrichment tests, we were able to examine the immune cell infiltration of the patients. There is a strong link between HCM and monocytic lineage, endothelial cells, natural killer T cells, memory B cells, immature B cells, ImmuneScore, and CD8 + T cells. However, more research is necessary to pinpoint the exact mechanisms involved. Finally, we found the ten most likely drugs that had targets by using targeted drug prediction for the blue module genes that were screened from the data. To forecast the binding ability, we molecularly docked each of these 10 potential targeted medicines with CEBPD. Abt-751 had the strongest binding ability, which suggests it could be used as a new HCM-targeting drug. We did not, however, confirm this with specific medications, which would have provided persuasive evidence for more research. Animals can initially receive these drugs intraperitoneally or by gavage to examine their impact on HCM. In vitro research can be done concurrently to look at the medications' actual mechanism of action. After that, we want to gather samples of human HCM to look at the impact of CEBPD on HCM development. Future studies on CEBPD can be conducted in humans and animals, respectively, to learn more about their distinct processes and the genetically related antagonists. This study could help discover how to prevent and treat CEBPD, as well as how to help doctors manage patients and predict their recovery. However, our current study still has many limitations:Dataset Heterogeneity: While integrating GSE36961, GSE1145, and GSE32453 improved statistical power, batch effects from different platforms (GPL15389 vs. GPL570) may introduce bias. Our normalization mitigated but did not fully eliminate platform-specific variances. The number of Samples in our validation dataset is relatively small, and we need to increase the number of patients to validate the depth sufficiently; Experimental Model Constraints: The TAC-induced mouse model and Ang II-treated cardiomyocytes primarily reflect pressure-overload hypertrophy, which may not fully recapitulate genetic HCM pathophysiology. Future studies should validate CEBPD in human HCM tissues and MYH7-mutant models; Translational Gaps: Although Abt-751 showed strong in silico binding to CEBPD, its efficacy and safety in vivo remain unverified. Additionally, the immune infiltration results (e.g., M2 macrophage dominance) were derived from bulk RNA-seq data, which lacks single-cell resolution to delineate tissue-specific immune subsets; Mechanistic Depth: While we identified CEBPD’s association with mitochondrial dysfunction, the precise molecular pathways (e.g., mitophagy) require further CRISPR-based functional assays. These limitations highlight opportunities for future work, including single-cell omics in human HCM samples and preclinical trials of Abt-751. Nevertheless, our integrative approach provides a foundational framework for understanding CEBPD’s role in HCM and its potential as a therapeutic target. In summary, we have screened genes associated to mitochondrial dysfunction from transcriptome level to protein level for the first time in HCM by combining bioinformatics, multiple machine learning, and protein network level screening techniques. Furthermore, in order to fully explore the potential immunological targets of HCM, we integrated a number of the most widely used immune infiltration assays to examine its immune microenvironment. In the end, we made models of cells and animals to test our ideas, predicted possible medicines, and used molecular docking to give us a clue in our search for medicines that target HCM. 5 Conclusion We used PPI networks and a variety of machine learning techniques to screen for the novel target marker (CEBPD) of HCM. We came to the conclusion that CEBPD may function in HCM through the monocytic lineage, endothelial cells, natural killer T cells, memory B cells, immature B cells, and CD8 + T cells when combined with immuno-infiltration analysis of the samples. The validation ladder's CEBPD expression matched that of the experimental ladder. In vitro and in vivo, CEBPD mRNA levels were also lowered. We also thought that the following ten drugs might be HCM targets: Aripiprazole, Tazarotene, Rimexolone, Medrysone, Adarotene, Abt-751, Dexketoprofen, Ciclesonide, Tolmetin, and Resiquimod. Finally, we found that Abt-751 has the highest binding affinity with CEBPD, indicating its potential as a targeted drug. Declarations Ethics statement Not applicable. Author contributions The paper has been reviewed and approved by all authors. JC, DX, and YL came up with the original concept for this research. ZW and RL oversaw the procedure and evaluated the final article. The data analysis and gathering were handled by ZC. The table and manuscript were created by XC, RH, and LL. Graphs with LX and SJ plotted. ZW examined the info. The study was planned and the text was revised by FL, JW, and ZS. Funding This study was supported by the Natural Science Foundation of China (Grant Nos.82070291). The funder had no role in the decision to publish or preparation of the manuscript. Acknowledgments We thank Feng-chun Lu, Jia-mao Wang and Zhong-gui Shan for technology supporting and composition instruction. Supplementary Materials The full contents of the supplement are available online. Ethics approval statement This study and included experimental procedures were approved by the institutional animal care and use committee of Xiamen university (approval no. XMULAC20200137). All animal housing and experiments were conducted in strict accordance with the institutional guidelines for care and use of laboratory animals. Patient data from the GEO database were used in this study. Consent for publication Not applicable. Competing interests The authors declare that they have no competing interests. Data availability statement The datasets supporting our findings are presented in the article. GSE36961, GSE1145 and GSE32453 were downloaded from the GEO database (https://www.ncbi.nlm.nih.gov/geo/). All analyzed data are included in this published article. The original data are available. Ethics approval and consent to participate This research utilized published studies and consortia that have made their summary statistics publicly available. All original studies included in this research have obtained approval from their respective ethical review boards, and participants have provided informed consent. It is important to note that no individual-level data was utilized in this study. As a result, no new ethical review board approval was necessary for this research. Competing interests There are no conficts of interest to declare. Disclosure statement No potential conflict of interest was reported by the authors. Clinical trial number Not applicable. Author details 1 The First Affiliated Hospital of Xiamen University, School of Medicine Xiamen University, NO.55, Zhenhai Road, Siming District, Xiamen, Fujian ,361003, China 2 Department of General Surgery, Fujian Medical University Union Hospital, No. 29 Xinquan Road, Fuzhou, 350001, China. 3 Department of Cardiac Surgery, Xiangan Hospital of Xiamen University, School of Medicine, Xiamen University, Xiamen, Fujian ,361100, China. References Nakamura, M. & Sadoshima, J. Mechanisms of physiological and pathological cardiac hypertrophy. Nat. Reviews Cardiol. 15 (7), 387–407 (2018). McKenna, W. J., Maron, B. J. & Thiene, G. Classification, Epidemiology, and Global Burden of Cardiomyopathies. Circul. Res. 121 (7), 722–730 (2017). Shimizu, I. & Minamino, T. Physiological and pathological cardiac hypertrophy. J. Mol. Cell. Cardiol. 97 , 245–262 (2016). Lyon, R. C., Zanella, F., Omens, J. H. & Sheikh, F. Mechanotransduction in Cardiac Hypertrophy and Failure. Circul. Res. 116 (8), 1462–1476 (2015). Weber, K. T., Clark, W. A., Janicki, J. S. & Shroff, S. G. Physiologic versus pathologic hypertrophy and the pressure-overloaded myocardium. J. Cardiovasc. Pharmacol. 10 (Suppl 6), S37–50 (1987). Chien, K. R., Knowlton, K. U., Zhu, H. & Chien, S. Regulation of cardiac gene expression during myocardial growth and hypertrophy: molecular studies of an adaptive physiologic response. FASEB journal: official publication Federation Am. Soc. Experimental Biology . 5 (15), 3037–3046 (1991). Yamamoto, S., Kita, S., Iyoda, T., Yamada, T. & Iwamoto, T. New Molecular Mechanisms for Cardiovascular Disease: Cardiac Hypertrophy and Cell-Volume Regulation. J. Pharmacol. Sci. 116 (4), 343–349 (2011). Bernardo, B. C., Weeks, K. L., Pretorius, L. & McMullen, J. R. Molecular distinction between physiological and pathological cardiac hypertrophy: Experimental findings and therapeutic strategies. Pharmacol. Ther. 128 (1), 191–227 (2010). Ashrafian, H., Redwood, C., Blair, E. & Watkins, H. Hypertrophic cardiomyopathy: a paradigm for myocardial energy depletion. Trends Genet. 19 (5), 263–268 (2003). Elliott, P. & McKenna, W. J. Hypertrophic cardiomyopathy. Lancet 363 (9424), 1881–1891 (2004). Okajima, Y., Tanabe, Y., Takayanagi, M. & Aotsuka, H. A follow up study of myocardial involvement in patients with mitochondrial encephalomyopathy, lactic acidosis, and stroke-like episodes (MELAS). Heart 80 (3), 292–295 (1998). Mudhar, H. S., Wagner, B. E. & Suvarna, S. K. Electron microscopy of myocardial tissue. A nine year review. J. Clin. Pathol. 54 (4), 321–325 (2001). Grasso, M. et al. The mitochondrial DNA mutation T12297C affects a highly conserved nucleotide of tRNA Leu(CUN) and is associated with dilated cardiomyopathy. Eur. J. Hum. Genet. 9 (4), 311–315 (2001). Robbins, R. C. et al. Cardiac transplantation for hypertrophic cardiomyopathy associated with Sengers syndrome. Ann. Thorac. Surg. 60 (5), 1425–1427 (1995). Lucas, D. T., Aryal, P., Szweda, L. I., Koch, W. J. & Leinwand, L. A. Alterations in mitochondrial function in a mouse model of hypertrophic cardiomyopathy. Am. J. Physiol. Heart Circ. Physiol. 284 (2), H575–H83 (2003). Tardiff, J. C. et al. Cardiac troponin T mutations result in allele-specific phenotypes in a mouse model for hypertrophic cardiomyopathy. J. Clin. Invest. 104 (4), 469–481 (1999). Unno, K. et al. Relation of functional and morphological changes in mitochondria to myocardial contractile and relaxation reserves in asymptomatic to mildly symptomatic patients with hypertrophic cardiomyopathy. Eur. Heart J. 30 (15), 1853–1862 (2009). Ranjbarvaziri, S. et al. Altered Cardiac Energetics and Mitochondrial Dysfunction in Hypertrophic Cardiomyopathy. Circulation 144 (21), 1714–1731 (2021). Wang, H. C. et al. Scientific discovery in the age of artificial intelligence. Nature 620 (7972), 47–60 (2023). Jiang, T., Gradus, J. L. & Rosellini, A. J. Supervised Machine Learning: A Brief Primer. Behav. Ther. 51 (5), 675–687 (2020). Liu, J., Li, J., Wang, H. & Yan, J. Application of deep learning in genomics. Sci. China Life Sci. 63 (12), 1860–1878 (2020). Kiselev, V. Y., Andrews, T. S. & Hemberg, M. Challenges in unsupervised clustering of single-cell RNA-seq data. Nat. Rev. Genet. 20 (5), 273–282 (2019). Tenekedjiev, K. I. & Nikolova, N. D. Applications of Monte Carlo simulation in modelling of biochemical processes [Book Chapter]2011. Faouzi, J. & Colliot, O. Classic machine learning methods arXiv. Liu, Y. J. et al. Incorporating machine learning and PPI networks to identify mitochondrial fission-related immune markers in abdominal aortic aneurysms. Heliyon 10 (7), e27989 (2024). Pickering, J. W. Machine learning for decision-making in cardiology: a narrative review to aid navigating the new landscape. Rev. Esp. Cardiol. 76 (8), 645–654 (2023). Feeny, A. K. et al. Artificial Intelligence and Machine Learning in Arrhythmias and Cardiac Electrophysiology. Circulation-Arrhythmia Electrophysiol. ; 13 (8). (2020). Johnson, K. W. et al. Artificial Intelligence in Cardiology. J. Am. Coll. Cardiol. 71 (23), 2668–2679 (2018). Costa-Silva, J., Domingues, D. & Lopes, F. M. RNA-Seq differential expression analysis: An extended review and a software tool. Plos One ; 12 (12). (2017). Su, X. X. et al. Bioinformatics and machine learning approaches reveal key genes and underlying molecular mechanisms of atherosclerosis: A review. Medicine ; 103 (31). (2024). Zhao, W. et al. Weighted Gene Coexpression Network Analysis: State of the Art. J. Biopharm. Stat. 20 (2), 281–300 (2010). Xu, M. M. et al. Identification and validation of immune and oxidative stress-related diagnostic markers for diabetic nephropathy by WGCNA and machine learning. Front. Immunol. ; 14 . (2023). Chicco, D. & Jurman, G. A brief survey of tools for genomic regions enrichment analysis. Front. Bioinf. ; 2 . (2022). Kanehisa, M., Furumichi, M., Sato, Y., Matsuura, Y. & Ishiguro-Watanabe, M. KEGG: biological systems database as a model of the real world. Nucleic Acids Res. 53 (D1), D672–d7 (2025). Yu, G., Wang, L. G., Han, Y. & He, Q. Y. clusterProfiler: an R package for comparing biological themes among gene clusters. Omics 16 (5), 284–287 (2012). Petitprez, F. et al. Transcriptomic analysis of the tumor microenvironment to guide prognosis and immunotherapies. Cancer Immunol. Immunotherapy . 67 (6), 981–988 (2018). Aran, D., Hu, Z. C. & Butte, A. J. xCell: digitally portraying the tissue cellular heterogeneity landscape. Genome Biol. ; 18 . (2017). Li, S. N. et al. Intratumoral microbial heterogeneity affected tumor immune microenvironment and determined clinical outcome of HBV-related HCC. Hepatology 78 (4), 1079–1091 (2023). Yu, H. Z. et al. Identification and analysis of key hypoxia- and immune-related genes in hypertrophic cardiomyopathy. Biol. Res. ; 56 (1). (2023). Chen, B., Khodadoust, M. S., Liu, C. L., Newman, A. M. & Alizadeh, A. A. Profiling Tumor Infiltrating Immune Cells with CIBERSORT. In: VonStechow L, editor. Cancer Systems Biology: Methods and Protocols. Methods in Molecular Biology. 17112018. pp. 243 – 59. Binson, V. A. et al. A Review of Machine Learning Algorithms for Biomedical Applications. Ann. Biomed. Eng. 52 (4), 1051–1066 (2024). Rosati, D. et al. Differential gene expression analysis pipelines and bioinformatic tools for the identification of specific biomarkers: A review. Comput. Struct. Biotechnol. J. 23 , 1154–1168 (2024). Arbet, J., McGue, M., Chatterjee, S. & Basu, S. Resampling-based tests for Lasso in genome-wide association studies. BMC Genet. ; 18 . (2017). Kim, Y., Hao, J., Mallavarapu, T., Park, J. & Kang, M. Hi-LASSO: High-Dimensional LASSO. IEEE Access. 7 , 44562–44573 (2019). Sanz, H., Valim, C., Vegas, E., Oller, J. M. & Reverter, F. SVM-RFE: selection and visualization of the most relevant features through non-linear kernels. BMC Bioinform. ; 19 . (2018). Kavitha, K. R., Rajendran, G. S. & Varsha, J. (eds) A Correlation Based SVM-Recursive Multiple Feature Elimination Classifier for Breast Cancer Disease using Microarray. International Conference on Advances in Computing, Communications and Informatics (ICACCI); Sep 21–24; Jaipur, INDIA2016. (2016). Díaz-Uriarte, R. & Alvarez de Andrés, S. Gene selection and classification of microarray data using random forest. BMC Bioinform. 7 , 3 (2006). Li, J., Malley, J. D., Andrew, A. S., Karagas, M. R. & Moore, J. H. Detecting gene-gene interactions using a permutation-based random forest method. BioData Min. ; 9 . (2016). Motsinger-Reif, A. A., Deodhar, S., Winham, S. J. & Hardison, N. E. Grammatical evolution decision trees for detecting gene-gene interactions. BioData Min. ; 3 . (2010). Yeh, J. Y. & Wu, T. H. Cascade of genetic algorithm and decision tree for cancer classification on gene expression data. Expert Syst. 27 (3), 201–218 (2010). Li, W., Yin, Y. B., Quan, X. W. & Zhang, H. Gene Expression Value Prediction Based on XGBoost Algorithm. Front. Genet. ; 10 . (2019). Yang, L. Z., Chen, Y. W. & Huang, W. Hub Genes Identification, Small Molecule Compounds Prediction for Atrial Fibrillation and Diagnostic Model Construction Based on XGBoost Algorithm. Front. Cardiovasc. Med. ; 9 . (2022). Guo, Y. J. et al. Gene-Based Testing of Interactions Using XGBoost in Genome-Wide Association Studies. Front. Cell. Dev. Biol. ; 9 . (2021). Wang, L. Z. et al. Transformer fault diagnosis method based on SMOTE and NGO-GBDT. Sci. Rep. ; 14 (1). (2024). Zhang, Y. et al. A feature selection and multi-model fusion-based approach of predicting air quality. ISA Trans. 100 , 210–220 (2020). Peng, C., Zhang, Y. X., Lang, X. Y. & Zhang, Y. Role of mitochondrial metabolic disorder and immune infiltration in diabetic cardiomyopathy: new insights from bioinformatics analysis. J. Translational Med. ; 21 (1). (2023). Zhao, Y. C., Chen, X. Q., Chen, J. J. & Qi, X. Decoding Connectivity Map-based drug repurposing for oncotherapy. Brief. Bioinform. ; 24 (3). (2023). Musa, A. et al. A review of connectivity map and computational approaches in pharmacogenomics. Brief. Bioinform. 19 (3), 506–523 (2018). Lamb, J. The Connectivity Map: a new tool for biomedical research. Nat. Rev. Cancer . 7 (1), 54–60 (2007). Ji, L. Y. et al. Identification of bioactive compounds and potential mechanisms of scutellariae radix-coptidis rhizoma in the treatment of atherosclerosis by integrating network pharmacology and experimental validation. Biomed. Pharmacother. ; 165 . (2023). Ommen, S. R. & Nishimura, R. A. Hypertrophic cardiomyopathy. Curr. Probl. Cardiol. 29 (5), 239–291 (2004). Ullmann, T., Luckhardt, S., Wolf, M., Parnham, M. J. & Resch, E. High-Throughput Screening for CEBPD-Modulating Compounds in THP-1-Derived Reporter Macrophages Identifies Anti-Inflammatory HDAC and BET Inhibitors. Int. J. Mol. Sci. ; 22 (6). (2021). Wedel, A. & Ziegler-Heitbrock, H. W. L. The C/EBP Family of Transcription Factors. Immunobiology 193 (2–4), 171–185 (1995). Akasaka, T. et al. Five members of the CEBP transcription factor family are targeted by recurrent IGH translocations in B-cell precursor acute lymphoblastic leukemia (BCP-ALL). Blood 109 (8), 3451–3461 (2007). Chan, T. C., Shiue, Y. L. & Li, C. F. The biological impacts of CEBPD on urothelial carcinoma development and progression. Front. Oncol. ; 13 . (2023). Spek, C. A., Aberson, H. L., Butler, J. M., de Vos, A. F. & Duitman, J. CEBPD Potentiates the Macrophage Inflammatory Response but CEBPD Knock-Out Macrophages Fail to Identify CEBPD-Dependent Pro-Inflammatory Transcriptional Programs. Cells ; 10 (9). (2021). Ko, C. Y., Chang, W. C. & Wang, J. M. Biological roles of CCAAT/Enhancer-binding protein delta during inflammation. J. Biomed. Sci. ; 22 . (2015). Ullmann, T., Luckhardt, S., Wolf, M., Parnham, M. J. & Resch, E. High-Throughput Screening for CEBPD-Modulating Compounds in THP-1-Derived Reporter Macrophages Identifies Anti-Inflammatory HDAC and BET Inhibitors. Int. J. Mol. Sci. ; 22 (6). (2021). Wang, J. M., Ko, C. Y., Chen, L. C., Wang, W. L. & Chang, W. C. Functional role of NF-IL6beta and its sumoylation and acetylation modifications in promoter activation of cyclooxygenase 2 gene. Nucleic Acids Res. 34 (1), 217–231 (2006). Cho, M. K., Cho, Y. H., Lee, G. H. & Kim, S. G. Induction of cyclooxygenase-2 by bovine type I collagen in macrophages via C/EBP and CREB activation by multiple cell signaling pathways. Biochem. Pharmacol. 67 (12), 2239–2250 (2004). Thomas, B. et al. Critical role of C/EBPdelta and C/EBPbeta factors in the stimulation of the cyclooxygenase-2 gene transcription by interleukin-1beta in articular chondrocytes. Eur. J. Biochem. 267 (23), 6798–6809 (2000). Balamurugan, K. et al. FBXW7α attenuates inflammatory signalling by downregulating C/EBPδ and its target gene Tlr4. Nat. Commun. ; 4 . (2013). Wang, Y. H. et al. CEBPD amplification and overexpression in urothelial carcinoma: a driver of tumor metastasis indicating adverse prognosis. Oncotarget 6 (31), 31069–31084 (2015). Tsai, H. H. et al. Metformin promotes apoptosis in hepatocellular carcinoma through the CEBPD-induced autophagy pathway. Oncotarget 8 (8), 13832–13845 (2017). Wang, Q. W. et al. A hierarchical and collaborative BRD4/CEBPD partnership governs vascular smooth muscle cell inflammation. Mol. Therapy-Methods Clin. Dev. 21 , 54–66 (2021). Chi, J. Y. et al. Fibroblast CEBPD/SDF4 axis in response to chemotherapy-induced angiogenesis through CXCR4. Cell. Death Discovery ; 7 (1). (2021). Chan, T. C. et al. Biological significance of MYC and CEBPD coamplification in urothelial carcinoma: Multilayered genomic, transcriptional and posttranscriptional positive feedback loops enhance oncogenic glycolysis. Clin. Transl Med. 11 (12), e674 (2021). Wang, W. J. et al. MCL1 participates in leptin-promoted mitochondrial fusion and contributes to drug resistance in gallbladder cancer. JCI Insight ; 6 (15). (2021). Banerjee, S. et al. Loss of C/EBPδ enhances IR-induced cell death by promoting oxidative stress and mitochondrial dysfunction. Free Radic Biol. Med. 99 , 296–307 (2016). Guha, M., Srinivasan, S., Koenigstein, A., Zaidi, M. & Avadhani, N. G. Enhanced osteoclastogenesis by mitochondrial retrograde signaling through transcriptional activation of the cathepsin K gene. Ann. N Y Acad. Sci. 1364 (1), 52–61 (2016). Löffler, M. C. et al. Protein kinase D1 deletion in adipocytes enhances energy dissipation and protects against adiposity. Embo j. ; 37 (22). (2018). Ommen, S. R. et al. 2020 AHA/ACC Guideline for the Diagnosis and Treatment of Patients With Hypertrophic Cardiomyopathy: Executive Summary A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. J. Am. Coll. Cardiol. 76 (25), 3022–3055 (2020). Maron, B. J., Maron, M. S., Maron, B. A. & Loscalzo, J. Moving Beyond the Sarcomere to Explain Heterogeneity in Hypertrophic Cardiomyopathy JACC Review Topic of the Week. J. Am. Coll. Cardiol. 73 (15), 1978–1986 (2019). Carrillo-Salinas, F. J., Ngwenyama, N., Anastasiou, M., Kaur, K. & Alcaide, P. Heart Inflammation Immune Cell Roles and Roads to the Heart. Am. J. Pathol. 189 (8), 1482–1494 (2019). Ozyilmaz, S. et al. The importance of the neutrophil-to-lymphocyte ratio in patients with hypertrophic cardiomyopathy. Rev. Port. Cardiol. 36 (4), 239–246 (2017). Zheng, X. F., Yang, Y., Fu, C. H. & Huang, R. A. Identification and verification of promising diagnostic biomarkers in patients with hypertrophic cardiomyopathy associate with immune cell infiltration characteristics. Life Sci. ; 285 . (2021). Additional Declarations No competing interests reported. Supplementary Files SupplementaryTable1.xlsx SupplementaryTable2.xlsx SupplementaryTable3.xlsx SupplementaryFigure1.tif SupplementaryFigure2.tif Westernblot.pdf ARRIVEGuidline.tif SupplementaryMaterial1.docx SupplementaryMaterialsLegenddescription.docx Cite Share Download PDF Status: Published Journal Publication published 29 Apr, 2025 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Accepted 04 Apr, 2025 Reviews received at journal 04 Apr, 2025 Reviews received at journal 31 Mar, 2025 Reviewers agreed at journal 24 Mar, 2025 Reviewers agreed at journal 21 Mar, 2025 Reviewers invited by journal 21 Mar, 2025 Submission checks completed at journal 21 Mar, 2025 First submitted to journal 14 Mar, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5122992","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":432185981,"identity":"58b9dc93-cb5d-4ffb-99df-66617ca42985","order_by":0,"name":"Jia-lin Chen","email":"","orcid":"","institution":"First Affiliated Hospital of Xiamen University","correspondingAuthor":false,"prefix":"","firstName":"Jia-lin","middleName":"","lastName":"Chen","suffix":""},{"id":432185982,"identity":"5d9940d5-2497-47d2-948c-e009074acc8e","order_by":1,"name":"Di Xiao","email":"","orcid":"","institution":"First Affiliated Hospital of Xiamen University","correspondingAuthor":false,"prefix":"","firstName":"Di","middleName":"","lastName":"Xiao","suffix":""},{"id":432185983,"identity":"a6c49e86-337d-43bb-9c53-01d83aca0ba1","order_by":2,"name":"Yi-jiang Liu","email":"","orcid":"","institution":"First Affiliated Hospital of Xiamen University","correspondingAuthor":false,"prefix":"","firstName":"Yi-jiang","middleName":"","lastName":"Liu","suffix":""},{"id":432185984,"identity":"e15a8095-2cc6-48bf-9e09-a343ad4a690e","order_by":3,"name":"Zhan Wang","email":"","orcid":"","institution":"First Affiliated Hospital of Xiamen University","correspondingAuthor":false,"prefix":"","firstName":"Zhan","middleName":"","lastName":"Wang","suffix":""},{"id":432185985,"identity":"1eb5eebe-5ce3-4c52-96da-99f2996ac9cc","order_by":4,"name":"Zhi-huang Chen","email":"","orcid":"","institution":"First Affiliated Hospital of Xiamen University","correspondingAuthor":false,"prefix":"","firstName":"Zhi-huang","middleName":"","lastName":"Chen","suffix":""},{"id":432185986,"identity":"7e50a474-1a48-4200-8a36-96b42e9fd25c","order_by":5,"name":"Rui Li","email":"","orcid":"","institution":"First Affiliated Hospital of Xiamen University","correspondingAuthor":false,"prefix":"","firstName":"Rui","middleName":"","lastName":"Li","suffix":""},{"id":432185987,"identity":"fdc35a60-4ad6-4172-8f8f-39ccf4e821b4","order_by":6,"name":"Li Li","email":"","orcid":"","institution":"First Affiliated Hospital of Xiamen University","correspondingAuthor":false,"prefix":"","firstName":"Li","middleName":"","lastName":"Li","suffix":""},{"id":432185988,"identity":"c13d7518-ffa7-4926-bcfd-a7657446f4cd","order_by":7,"name":"Rong-hai He","email":"","orcid":"","institution":"Xiangan Hospital of Xiamen University","correspondingAuthor":false,"prefix":"","firstName":"Rong-hai","middleName":"","lastName":"He","suffix":""},{"id":432185989,"identity":"8741463a-79b5-4ee1-a08a-903d378cb47c","order_by":8,"name":"Shu-yan Jiang","email":"","orcid":"","institution":"First Affiliated Hospital of Xiamen University","correspondingAuthor":false,"prefix":"","firstName":"Shu-yan","middleName":"","lastName":"Jiang","suffix":""},{"id":432185990,"identity":"4b05867f-bd68-4b2d-bac6-dd833636dc56","order_by":9,"name":"Xin Chen","email":"","orcid":"","institution":"First Affiliated Hospital of Xiamen University","correspondingAuthor":false,"prefix":"","firstName":"Xin","middleName":"","lastName":"Chen","suffix":""},{"id":432185991,"identity":"034f93ae-e7fc-44ee-9334-e6911a04d8b5","order_by":10,"name":"Lin-xi Xu","email":"","orcid":"","institution":"First Affiliated Hospital of Xiamen University","correspondingAuthor":false,"prefix":"","firstName":"Lin-xi","middleName":"","lastName":"Xu","suffix":""},{"id":432185992,"identity":"f7767dc0-0d74-494c-b2a0-ad2e7ae55114","order_by":11,"name":"Feng-chun Lu","email":"","orcid":"","institution":"Fujian Medical University Union Hospital","correspondingAuthor":false,"prefix":"","firstName":"Feng-chun","middleName":"","lastName":"Lu","suffix":""},{"id":432185993,"identity":"a65e5b2e-b131-44d6-af2c-8ddd48076202","order_by":12,"name":"Jia-mao Wang","email":"","orcid":"","institution":"First Affiliated Hospital of Xiamen University","correspondingAuthor":false,"prefix":"","firstName":"Jia-mao","middleName":"","lastName":"Wang","suffix":""},{"id":432185995,"identity":"6d435243-08a0-4da5-9af5-409f7cba0c82","order_by":13,"name":"Zhong-gui Shan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAxElEQVRIiWNgGAWjYJACCTDJ3tj44AMJWgwYGHgONxvOIE2LRHqbNAcxyuX7j1+88aHmj5z5zIcN0gwMdnK6DQS0GBw4U2w545iBscztxAbjAoZkY7MDhLQw9qRJ8zYYJM6QTmxInsFwIHEbIS3yzTxQLZIHGw7zEKOF4Rj7MYgWCcbGZqK0GJzhYQb6xdhYgiexmXGGARF+AYbYQ2CIyclJsB9//uNDhZ0cQS3AKDRAtpSgchBgf0CUslEwCkbBKBjBAADc2kBGh3d1+QAAAABJRU5ErkJggg==","orcid":"","institution":"First Affiliated Hospital of Xiamen University","correspondingAuthor":true,"prefix":"","firstName":"Zhong-gui","middleName":"","lastName":"Shan","suffix":""}],"badges":[],"createdAt":"2024-09-20 10:45:58","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5122992/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5122992/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-025-97534-x","type":"published","date":"2025-04-29T15:57:05+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":79089705,"identity":"812642b5-e024-4f23-84f3-71df50a77070","added_by":"auto","created_at":"2025-03-24 09:46:36","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1738866,"visible":true,"origin":"","legend":"\u003cp\u003eThe technical route of this article.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/3ea9358b2ae4ddb75b3af78f.png"},{"id":79089713,"identity":"13feb311-12b3-444e-919a-e238e733e959","added_by":"auto","created_at":"2025-03-24 09:46:37","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1674616,"visible":true,"origin":"","legend":"\u003cp\u003eDifferential Expression Analysis and weighted gene co-expression network analysis (WGCNA) of different expression genes (DEGs)\u003cstrong\u003e. \u003c/strong\u003e(A) Based on differential expression analysis, a volcano map displays down-regulated genes (blue), no significant genes (grey), and up-regulated genes (red). (B) A heatmap that clusters the DEGs of two groups and displays how their expressions differ from one another. (C) Gathering samples and organizing them into several clinical categories. (D) The relationship between the scale-free fit index and various soft-thresholding powers. (E) The relationship between the mean connection and various soft-thresholding capabilities. (F) Unique modules are indicated by different colors in the gene clustering tree diagram. (G) Assembling clusters according to different values of module features. (H) The clinical characteristics and the relationships among the 14 modules. (I) The scatterplot in the blue module that shows how membership (MM) and gene significance (GS)are related.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/b9b2073efeaa9f749408c6d3.png"},{"id":79089739,"identity":"f2ed35e0-41c9-459f-8697-d0448e21b79e","added_by":"auto","created_at":"2025-03-24 09:46:39","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1512831,"visible":true,"origin":"","legend":"\u003cp\u003eFunctional enrichment analysis. (A) The Gene Ontology (GO) enrichment results for the blue module, broken down into molecular functions (MF), cellular components (CC), and biological processes (BP). (B, C) The blue module illustrates the enrichment of the Kyoto Encyclopedia of Genomes (KEGG) pathway. (D) Samples with Gene Set Variation Analysis (GSVA) are displayed.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/597991ffb7caedecd56787b9.png"},{"id":79089711,"identity":"3b0c278b-5b13-4e22-9d6a-6adefa455972","added_by":"auto","created_at":"2025-03-24 09:46:37","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":195111,"visible":true,"origin":"","legend":"\u003cp\u003eThe immune microenvironment of HCM is the focus of this study. (A-C) We utilized an xCell analysis box plot to display the immunization ratings of the aggregated samples. We used ESTIMATE analysis to find the stromal score, the immune score, and the estimate score for HCM samples. (E) We analyzed the samples for HCM using the EPIC immunoassay. (F) MCPcounter research revealed that 10 immune-related cells were increased in the sample. (G) A box plot from the CIBERSORTX analysis shows that the sample had an enrichment of 22 immune-related cells. (H-K) The heat map and box plot were produced by the ssGSEA analysis of the samples, respectively. We used the rank sum test to analyze the data. The data were shown as the mean ± SD for all n = 3. *p \u0026lt; 0.05, **p \u0026lt; 0.01, ***p \u0026lt; 0.001, and ****p \u0026lt; 0.0001 vs. the control group.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/4dfa3229a609a621743ad0b4.png"},{"id":79089730,"identity":"2c48f98b-67d4-4eb0-b207-28f944dbeca7","added_by":"auto","created_at":"2025-03-24 09:46:38","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":835969,"visible":true,"origin":"","legend":"\u003cp\u003eProtein-protein interaction (PPI) networks and machine learning work together to screen for the most characteristic gene (MCG). (A, B) Feature gene screening with the Lasso model. We identify non-zero coefficients in the optimal model by utilizing rainbow and binomial deviation plots. (C, D) IncMSE and IncNodePurity Random Forest Analysis is used to display the top variables. (E) A branch diagram showing decision trees filtering for characteristic variables. (F, G) The picture shows the features of the ten genes by using the support vector machine-recursive feature elimination (SVM-RFE) method. We can find the optimal n=6 for training RMSE by varying the quantity of variables. (H, I) Force and waterfall diagrams are used to display the SHapley Additive exPlanations (SHAP) values of the blue module genes, which were determined using the XGBoost model. (J) Waterfall charts illustrate the blue module genes' SHAP values, which were determined using the LightGBM model. (K, L) Bar graphs displaying the most highly characterized variables in the XGBoost and LightGBM models, respectively. (M) For visualization, blue module genes are imported into Cytoscape via the PPI network. (N) Cytohubba's top genes were calculated in Cytoscape. (O) The top genes in Cytoscape were determined using MCODE calculations. (P) Venn diagram showing the intersection of genes screened between lasso, random forest, SVM-REF, XGBoost, LightGBM, CytoHubba, MCODE, and mitochondrial dysfunction-related genes.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/0a6c64b5f038f85fac1dac7c.png"},{"id":79090982,"identity":"68c2cb6d-558b-4650-bb4d-a592e56f3e8b","added_by":"auto","created_at":"2025-03-24 09:54:37","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":47084,"visible":true,"origin":"","legend":"\u003cp\u003eValidation ladders of the most characteristic gene (MCG). (A) Transcriptomic data from three datasets, GSE32453, GSE35229, and GSE1145, were pooled into a validation ladder. (B) The validation ladder displays the expression levels of mRNA. (C, D) The ROC curves represent the MCG diagnostic value in the experimental and validation ladder, respectively. For the data statistics, the t-test was employed. Data were presented as the mean ± SD (n ≥ 3). *p \u0026lt; 0.05, **p \u0026lt; 0.01, ***p \u0026lt; 0.001, and ****p \u0026lt; 0.0001 vs. the Sham group or Control group.\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/3303af0dd9cfeaa791aa91a7.png"},{"id":79089718,"identity":"10c563ce-1fce-4dae-bcc2-b18f16f33ad7","added_by":"auto","created_at":"2025-03-24 09:46:37","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":3618201,"visible":true,"origin":"","legend":"\u003cp\u003eGross morphological changes in the HCM and neonatal mouse primary cardiomyocyte stimulation model. (A, B) Representative plots of mouse echocardiograms, LVEF, and LVFS statistics. (C) Typical diagrams of large mouse heart specimens. (D) Statistical Plot of HW/TL (mg/mm) and HW/BW (mg/g). (E) a-b are respectively HE staining, Masson staining, Picro-Sirius Red (PSR) stain, and WGA diagrams of representation (50um and 1000um). (F) The three statistical graphs count the area of Masson staining, Picro-Sirius Red (PSR) stain, and WGA and the size of the cells, respectively. (G, H) The Western blot demonstrated the expression of ANP and -MHC protein and mRNA levels in an animal model of HCM. (I and J) The levels of ANP and β-MHC protein and mRNA were examined in a cell model made of newborn mouse primary cardiomyocytes. For the data statistics, the t-test was employed. Data were presented as the mean ± SD (n ≥ 3). *p \u0026lt; 0.05, **p \u0026lt; 0.01, ***p \u0026lt; 0.001, and ****p \u0026lt; 0.0001 vs. the Sham group or Control group.\u003c/p\u003e","description":"","filename":"7.png","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/c63d94be0817b7bc8ce0fd27.png"},{"id":79089757,"identity":"618da21d-bc90-4ed1-a821-dc31dbd4516f","added_by":"auto","created_at":"2025-03-24 09:46:40","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":637517,"visible":true,"origin":"","legend":"\u003cp\u003eExpression of the most characteristic gene (MCG)in vivo and in vitro. (A, B) Representative graphs of TUNEL staining showing apoptosis of sections from TAC and control groups and corresponding statistics. (C, D). The mRNA level of CEBPD expression in vivo and in vitro is shown by the box plot. (E, F) The study used Western blotting to demonstrate the expression of CEBPD in animal tissues, along with the corresponding statistics. For the data statistics, the t-test was employed. Data were presented as the mean ± SD (n ≥ 3). *p \u0026lt; 0.05, **p \u0026lt; 0.01, ***p \u0026lt; 0.001, and ****p \u0026lt; 0.0001 vs. the Sham group or Control group.\u003c/p\u003e","description":"","filename":"8.png","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/f10b50f0548e186118a5e0c7.png"},{"id":79089767,"identity":"be5947e9-3497-4337-966b-70d4dee037c0","added_by":"auto","created_at":"2025-03-24 09:46:40","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":2326410,"visible":true,"origin":"","legend":"\u003cp\u003ePrediction of targeting drugs for HCM by Cmap. (A) The graph plots the RAW cs, Norm cs, doses, times, N-samples, and fdr_q_nlog10 of the Top 10 targeted drugs. (B-K) Each of the ten molecular docking diagrams represents the binding of Adarotene, Aripiprazole, Ciclesonide, Abt-751, Dexketoprofen, Medrysone, Resiquimod, Rimexolone, Tazarotene, and Tolmetin to the most characteristic gene (CEBPD). (L) The bars represent the binding strength of the ten drug monomers to CEBPD.\u003c/p\u003e","description":"","filename":"9.png","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/b9bff13223d7810f09321355.png"},{"id":81987456,"identity":"daf33b12-d151-4900-ac1e-47c0e723a6ee","added_by":"auto","created_at":"2025-05-05 16:03:20","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":13347556,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/408a643a-cf7a-4690-bb73-5ec5a4238550.pdf"},{"id":79089729,"identity":"c2bf020e-0f13-494e-9c75-e6fc04bb8c56","added_by":"auto","created_at":"2025-03-24 09:46:38","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":644987,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/abed325a400254aded619bf0.xlsx"},{"id":79089707,"identity":"0eca5a3f-2348-4f23-88a3-044f9eb9e572","added_by":"auto","created_at":"2025-03-24 09:46:37","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":9667,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/74e7453d398a3c29363144a7.xlsx"},{"id":79090988,"identity":"becc4307-a566-454e-b3c4-65244f8c5cd3","added_by":"auto","created_at":"2025-03-24 09:54:38","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":47583,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable3.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/dc9ce7d54e60ea6a996d844b.xlsx"},{"id":79089792,"identity":"cf9733c7-2bf9-4239-9f46-5536913b6421","added_by":"auto","created_at":"2025-03-24 09:46:42","extension":"tif","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":68877280,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure1.tif","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/47c478066557219653cb2b68.tif"},{"id":79089770,"identity":"eff031c8-c036-4d51-bf77-9809a6cce4ee","added_by":"auto","created_at":"2025-03-24 09:46:41","extension":"tif","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":20481140,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure2.tif","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/7bf5b1d53551b377e6f3641c.tif"},{"id":79090981,"identity":"7e1a9d46-594b-4e25-914e-7510e3a665fb","added_by":"auto","created_at":"2025-03-24 09:54:36","extension":"pdf","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":508943,"visible":true,"origin":"","legend":"","description":"","filename":"Westernblot.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/ed198b298a12851912b3dc49.pdf"},{"id":79089708,"identity":"58016d47-7857-474b-8ad1-169184c3dc40","added_by":"auto","created_at":"2025-03-24 09:46:37","extension":"tif","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":285378,"visible":true,"origin":"","legend":"","description":"","filename":"ARRIVEGuidline.tif","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/11a0e666da42de41d3b623ec.tif"},{"id":79090985,"identity":"13687c22-b055-4096-85a6-7a688f1a3cae","added_by":"auto","created_at":"2025-03-24 09:54:37","extension":"docx","order_by":8,"title":"","display":"","copyAsset":false,"role":"supplement","size":46410,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterial1.docx","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/078804fd63f425e83b4661f2.docx"},{"id":79090992,"identity":"ebca7826-ab45-4940-9ffa-f2e54b01ddce","added_by":"auto","created_at":"2025-03-24 09:54:39","extension":"docx","order_by":9,"title":"","display":"","copyAsset":false,"role":"supplement","size":15108,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterialsLegenddescription.docx","url":"https://assets-eu.researchsquare.com/files/rs-5122992/v1/b7403d3edc6ac0aee5e3f00c.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Protein interactions, network pharmacology, and machine learning work together to predict genes linked to mitochondrial dysfunction in hypertrophic cardiomyopathy","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003eHeart mass and volume can go up because of cardiomyocyte hypertrophy, myocardial interstitial cell hyperplasia, changes in the extracellular matrix, and other things. This condition is known as hypertrophic cardiomyopathy (HCM), and it is a major predictor of the course and prognosis of cardiac disease [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. It also plays a key role in cardiac remodeling and is an independent risk factor for cardiac events. Pathological and physiological myocardial hypertrophy are two types of cardiac hypertrophy [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Pathological cardiac hypertrophy causes unfavorable cardiovascular events such as heart failure, malignant arrhythmias, and sudden death among these patients. It is thought that between 1 in 200 and 1 in 500 people have pathologic cardiomyopathy. This is because more clinical and molecular genetic studies are being done, especially since more sensitive diagnostic cardiac imaging and family tree screening have become popular [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe majority of cardiomyocytes do not proliferate under physiological settings and instead undergo terminal differentiation in maturity. The heart can adapt to environmental demands thanks to the plasticity of cardiac tissue, and cells can grow, shrink, or die in response to a range of physiologic or pathologic stimuli [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Initially, hypertrophy is an adaptive reaction to pathological and physiological stressors. In order for the adult heart to keep working, there is a decrease in ventricular wall stress. The type and level of this decrease depend on the stimulus's type, intensity, and length [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. This is due to an increase in the size of individual cardiomyocytes rather than an increase in their number. Physiological hypertrophy, which is frequently observed with exercise, pregnancy, etc., is characterized by normal or improved contractile function [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eA lot of different neurohumoral mediators, hemodynamic overload, and other factors do affect the heart's ability to work normally in heart valve disease, cardiomyopathy, ischemic heart disease, high blood pressure, and other conditions [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. To deal with the extra stress on the heart, cardiomyocytes grow faster than the capillaries can supply enough oxygen and nutrients. This causes cardiac myocyte proteins to be made more quickly, cardiac muscle cells to get bigger, and interstitial fibrosis of cardiomyocytes [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Heart systolic and diastolic functions get better, heart valve compliance gets worse, and more oxygen is used because cardiomyocytes make more proteins, have more volume, and have more interstitial fibrosis and intercellular stroma. This leads to pathological myocardial hypertrophy, which subsequently progresses to heart failure, malignant arrhythmia, and sudden cardiac death [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eNevertheless, there is still room for improvement in the management and prevention of pathological ventricular hypertrophy in many clinical presentations. In order to halt the course of pathological cardiac hypertrophy, it is crucial to understand the regulatory processes underlying the condition and identify useful treatment targets.\u003c/p\u003e \u003cp\u003eMost cases of pathological cardiac hypertrophy lead to higher overall adenosine triphosphate (ATP) production and use, which increases the body's need for energy [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Patients with mutations in their mitochondrial DNA typically develop pathologic cardiac hypertrophy. This suggests that pathologic changes in pathologic cardiac hypertrophy are caused by changed mitochondrial state [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. According to electron microscopy findings, during the latter phases of myocardial hypertrophy, mitochondrial fission and cristae disruption are crucial for cardiac remodeling [\u003cspan additionalcitationids=\"CR13\" citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Furthermore, Lucas and Tardiff found that patients' heart tissue and a mouse model with HCM had impaired mitochondrial architecture and function [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. The molecular mechanisms underlying the onset and progression of pathological cardiac hypertrophy, as well as changes in mitochondrial structure and function, are still largely unknown, despite the fact that current evidence suggests that pathological cardiac hypertrophy is related to energy metabolism, specifically mitochondrial function [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eBy combining and altering the original fundamental algorithms, machine learning (ML), which has its roots in computer science and mathematics, suggests a new set of algorithms for building inference and predictive data models [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. Supervised machine learning has the greatest variety of uses. Regression and clustering are further tasks in machine learning [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. The learning function in regression problems assigns a real value to each piece of data. After that, the value of the predictor variable can be determined for every new sample by using the estimates from earlier samples [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. A popular unsupervised problem is clustering, where the goal is to identify one or more characteristics to classify in order to explain the data set. Every new sample in this procedure can be categorized into one of the recognized clusters according to the shared characteristics that they have. The training set and the test set are the two distinct sets created from the data samples [\u003cspan citationid=\"CR87\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. Next, the training set is used to construct the model, and the test set is used to estimate the model's performance [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. We can utilize a variety of machine learning techniques after preprocessing the data and determining the kind of learning task. Artificial neural networks, support vector machines, logistic and linear regression, and tree-based techniques are examples of machine learning algorithms that are often utilized in practice [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. Afterwards, these separate models can be integrated using ensemble learning, a technique that maximizes overall performance by utilizing several weak classifiers [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eBecause so many diagnostic and treatment choices in cardiology depend on digital patient-specific data, such as electrocardiograms (ECGs), echocardiograms, etc., and because medical knowledge is becoming more and more sophisticated, machine learning has a huge potential influence in the field [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. The volume of information available about healthcare is astounding and includes imaging, prescription lists, wearable technology and sensor data, clinical notes, and much more. AI algorithms can be used in a variety of ways as they become more prevalent in the field of clinical cardiology [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIn this study, we create in vitro and in vivo models to validate the most described target genes while also utilizing bioinformatics and machine learning techniques. Lastly, we forecast the matching targeted medications and validate them appropriately.\u003c/p\u003e"},{"header":"2 Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Subject\u003c/h2\u003e \u003cp\u003eThe Department of School of Medicine at Xiamen University conducted the experiments. Experiments were approved by the Institutional Animal Care and Use Committee at Xiamen University and conducted in accordance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals and with the ARRIVE guidelines. We made efforts to minimize animal pain and discomfort and to reduce the number of animals used. The C57BL/6 wild-type mice utilized in this study were kept in the SPF-grade environment of the Xiamen University Animal Experiment Center, which was acquired from Beijing Vital River. The body weight was kept between 22 and 25 grams. We weighed the mice and administered an intraperitoneal injection of sodium pentobarbital at a dosage of 50 mg/kg to induce anesthesia. The 12 mice that completed the experiment were 12 male C57BL/6J mice. Mice were maintained in a temperature-controlled environment (21\u0026thinsp;\u0026plusmn;\u0026thinsp;1\u0026deg;C) on a 12:12 light:dark cycle (lights on at 0600). Mice had free access to food and water throughout the experiment except for the brief time in the testing apparatus. Prior to entering the experiment, mice were group housed with mice from the same sex and same strain. Following TAC surgery, mice were singly housed for the remainder of the experiment. Mice were housed in cages that contained Nestlets and Shepherd Shacks both prior to the experiment when group housed and during the experiment when singly housed.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Patients and datasets from Gene Expression Omnibus (GEO) and Differential Expression Analysis\u003c/h2\u003e \u003cp\u003eThe Gene Expression Omnibus (GEO) database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ncbi.nlm.nih.gov/geo/\u003c/span\u003e\u003cspan address=\"https://www.ncbi.nlm.nih.gov/geo/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) was used to get clinical data and transcriptome expression profile matrices from people who had pathologic cardiac hypertrophy. The study used GSE36961(GPL15389) as the experimental ladder and had 106 people with hypertrophic cardiomyopathy (HCM) and 39 healthy controls. The GPL15389 platform was used to generate the data. We combined datasets GSE1145 (GPL570 platform) and GSE32453 (GPL6104 platform) to make a validation cohort with 13 HCM patients and 5 normal subjects for the validation ladder. All analyses were run in RStudio software (Version: 2024.04.0\u0026thinsp;+\u0026thinsp;735.pro3). Using the R package \"limma\"(Version:3.56.2) we normalized the GSE36961 expression profile matrix data and carried out a differential expression analysis. Differentially expressed genes (DEGs) were considered to be significantly differentially expressed when their adjusted p-value (FDR) was less than 0.05 and |log2FC| was greater than 1.0. [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Weighted gene co-expression network analysis of DEGs\u003c/h2\u003e \u003cp\u003eOne technique for network analysis that can be used to examine the relationships between genes is Weighted Gene Co-expression Network Analysis (WGCNA) [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. Using the R packages \"WGCNA\"(Version:1.73),\"reshape2\"(Version:1.4.4) and \"stringr\"(Version:1.5.0), we recreated the expression profile matrix containing only different expression gens (DEGs). We next determined the median absolute deviation of gene expression in each sample and removed the top 50% of genes with the smallest median absolute deviation [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. The choice of β is crucial in Weighted Gene Co-expression Network Analysis (WGCNA), as it determines the extent to which strong correlations are emphasized while weak correlations are suppressed. The β value was selected based on the principle of scale-free network topology, a fundamental property observed in biological networks where a few nodes (hubs) exhibit a high degree of connectivity while most nodes have fewer connections. To determine the optimal β, we analyzed the relationship between β and the scale-free topology fit index (R\u0026sup2;). The goal was to identify the smallest β at which R\u0026sup2; reaches a plateau, indicating that the network structure sufficiently approximates a scale-free topology. In this study, we incrementally increased β and calculated the corresponding R\u0026sup2; values. When β\u0026thinsp;=\u0026thinsp;10, the scale-free topology fit index reached R\u0026sup2; = 0.86, indicating that the network's connectivity pattern closely followed a scale-free distribution. Using a lower β might have resulted in an insufficient emphasis on strong connections, leading to the inclusion of noisy or spurious correlations. Conversely, an excessively high β could have over-penalized weak connections, potentially eliminating biologically meaningful interactions. Thus, β\u0026thinsp;=\u0026thinsp;10 was chosen as the optimal threshold to balance network sparsity and biological relevance while ensuring robust module detection.The adjacency matrix was turned into a Topological Overlap Matrix (TOM) with a soft thresholding power of β\u0026thinsp;=\u0026thinsp;10 to make the network more stable by reducing unwanted correlations and background noise. This parameter was selected as the minimum power value at which the scale-free topology fit index (R\u0026sup2;) reached a plateau (R\u0026sup2;=0.86), ensuring optimal network connectivity while preserving biological relevance. Subsequently, the topological dissimilarity measure (1\u0026thinsp;\u0026minus;\u0026thinsp;TOM) was calculated to facilitate downstream module detection through hierarchical clustering. To balance the granularity of module detection with biological interpretability, we configured the hierarchical clustering parameters as follows: Dynamic tree-cutting sensitivity was set to 3 to allow moderate subdivision of gene clusters, enabling detection of functionally refined modules while avoiding excessive fragmentation. Minimum module size was constrained to 30 genes, excluding small clusters that likely represent random noise rather than biologically meaningful co-expression patterns. Module merging threshold was defined at a topological dissimilarity of 0.25 (equivalent to 1\u0026thinsp;\u0026minus;\u0026thinsp;TOM\u0026thinsp;\u0026ge;\u0026thinsp;0.75), ensuring consolidated modules maintain distinct expression profiles (Pearson correlation r\u0026thinsp;\u0026lt;\u0026thinsp;0.75) while preserving functional coherence. This parameter combination achieved optimal modular resolution where: Higher sensitivity (deepSplit\u0026thinsp;=\u0026thinsp;3) captured subtle expression variations Size filtering (minModuleSize\u0026thinsp;=\u0026thinsp;30) enhanced module reliability (GO enrichment p\u0026thinsp;\u0026lt;\u0026thinsp;1e\u003csup\u003e\u0026minus;\u0026thinsp;5\u003c/sup\u003e) Dissimilarity cutoff (cutHeight\u0026thinsp;=\u0026thinsp;0.25) prevented redundant modules (module eigengene correlation\u0026thinsp;\u0026lt;\u0026thinsp;0.85). So, we set the sensitivity to 3 and the minimum (gene group) of the gene tree to 30. Based on the computed values, the genes are clustered into distinct modular gene groups when the distance is less than 0.25. The module groups contain genes that exhibit significant linkage. A total of 14 modules were screened. Next, for every Module Membership (MM), we computed the Module Eigengene E (ME). Given the model's first main part, ME stands for the gene expression profile of the whole module. This is used to describe how the module's expression pattern changes in each sample. The correlation coefficient (cc) between a specific gene and a specific ME is presented. It's employed to characterize the dependability of genes associated with a specific module. To separate the modular genes from various modules, the weight threshold is set to 0.1, the MM threshold to 0.8, and the Gene Significance (GS) threshold to 0.1 [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Functional and pathway enrichment analysis\u003c/h2\u003e \u003cp\u003eGene Ontology (GO), Kyoto Encyclopedia of Genomes (KEGG) pathway, and Gene Set Variation Analysis (GSVA) are examples of functional enrichment analysis [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. Using the R package \"ClusterProfiler\"(Version:4.8.3), we carried out functional enrichment analysis of the Hub genes in the target modules in order to look into the possible biological roles and location of the chosen module genes in the organism [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e].We investigated the molecular functions (MF), cellular components (CC), and biological processes (BP) that the hub genes focused on through GO analysis. We used KEGG pathway analysis to find out what the hub genes do, mostly in organisms, so we could understand how genes interact in biological systems. We ranked genes by how important they were based on the enrichment scores we got from GSVA for each sample in the patient transcriptome expression matrix. This helped us predict the relevant pathways and biological mechanisms.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Analysis of the immune microenvironment\u003c/h2\u003e \u003cp\u003eWe used xCell, ESTIMATE, Estimate the Proportion of Immune and Cancer Cells (EPIC), Microenvironment Cell Populations-Counter (MCP-Counter), CIBERSORT, and Single Sample Genomic Enrichment Analysis (ssGSEA) methods to thoroughly compare the abundance of immune cells in the immune microenvironment of patients with pathologic cardiac hypertrophy with those of normal patients in order to investigate the changes in immune cells infiltrating around myocardial tissue in pathological cardiac hypertrophy [\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. xCell is implemented using the R package \"xCell\"(Version: 1.1.0) and is based on the ssGSEA algorithm for the abundance of 64 immune-associated cells [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. Using the stromal score, immune score, and estimate score\u0026mdash;all of which are included in the R package \"ESTIMATE\" (Version: 1.0.13)\u0026mdash;ESTIMATE assesses and contrasts the abundance of the appropriate cell types in the cohort samples. By using the R package \"EPIC\"(Version: 1.1,7),\"MCP-Counter\" (Version: 1.2.0) to calculate the geometric mean of marker gene expression, least squares regression explicitly introduces non-negativity constraints into the inverse fold product problem to compute the abundance of six immune cell types, fibroblasts, endothelial cells, and uncharacterized cells [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]. This allows for the quantification of the absolute abundance of eight immune cell types and two stromal cells. Using the concept of linear support vector regression, CIBERSORT (Version: 1.06) calculates the abundance of immune cells by de-convolution of the expression matrix of immune cell subtypes [\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]. To demonstrate the variations in immune cells among samples, ssGSEA(Version: 1.48.0) was carried out by computing the enriched scores (ES) of the 28 representative genes of the immune cells. variations in the samples' immune cells.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.6 Machine learning screening of feature gene\u003c/h2\u003e \u003cp\u003eHub genes can be screened to find feature genes using machine learning methods. Feature genes are a class of genes that are important in particular biological processes [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e]. This study used six machine learning algorithms: Random Forest (RF), DecisionTree, XGBoost, Support Vector Machine Recursive Feature Elimination (SVM-RFE), and Least Absolute Shrinkage and Selection Operator Regression (LASSO regression). The accuracy and efficiency of feature gene prediction using machine learning algorithms have significantly increased with the advancement of computational capacity and ongoing algorithmic development [\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e]. Regression analysis algorithms like LASSO regression can effectively filter out genes that significantly affect the anticipated target variables. These algorithms can be utilized to predict genes through the usage of R packages (\"glmnet\", Version: 4.1-7) [\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e]. The benefit of LASSO is that the variable selection increases the model's comprehensibility, while the regularization term limits the model's complexity and guards against overfitting [\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e]. The R package \"e1071\" (Version: 1.7\u0026ndash;13)can be used to implement Support Vector Machine Recursive Feature Elimination (SVM-RFE), a classification algorithm that uses 10-fold cross-validation and randomized repetitive cross-validation to extract the genes with higher weights of the variables and better generalization ability [\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e]. SVM-RFE establishes a threshold between two categories and identifies the most predictive feature genes based on the predicted feature vectors [\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e]. One technique for extracting genes with greater variable weights and improved generalization capacity is Random Forest (RF). The R software package \"RandomForest\" (Version: 4.1\u0026ndash;1.1) can be used to implement Random Forest (RF), an integrated learning technique that can predict continuous variables with virtually no substantial volatility in the prediction results and no constraint on continuous variables [\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e]. RF uses IncMSE and IncNodePurity to evaluate the genes; IncNodePurity is a better indicator of how genes affect the performance of the model. IncNodePurity is favored over IncMSE in RF's evaluation of genes since it more accurately captures the impact of genes on model performance [\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e]. To realize the most informative features that DecisionTree selects, use the R package \"rpart\" (Version: 4.1.19). Decision Tree is an intuitive tool for classification and regression that creates a model by recursively partitioning the dataset into smaller and smaller subsets and evaluating the information gain of each feature [\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e]. In addition to being a highly interpretable and simple to understand tool, decision trees can also be used as an integrated learning algorithm in other machine learning algorithms like Random Forests or Gradient Boosted Decision Trees (e.g., XGBoost). Decision trees can be used to identify gene combinations that have a significant impact on a given phenotype [\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e]. The R software package can be used to implement XGBoost, an integrated learning technique based on Gradient Boosted Decision Trees [\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e]. The R package \"XGBoost\" (Version: 1.7.5.1) implements an integrated learning method called XGBoost, which is based on gradient boosting decision trees. Based on feature importance scores, XGBoost is highly flexible and scalable, effectively handles missing values in genetic data and prevents overfitting, enhances prediction accuracy by combining multiple weak prediction models into a strong prediction model, and performs well in parallel computation [\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e]. As one of the algorithms in the Gradient Boosted Decision Trees (GBDT) framework and the subsequent creator of the XGBoost method, Light Gradient Boosting Machine (LightGBM) is an effective and scalable machine learning technique based on GBDT. Being a latecomer to the XGBoost algorithm but still a part of the GBDT framework, LightGBM builds upon the strengths of its predecessor algorithms, including XGBoost, and applies further optimizations to them by using The R package \"lightgbm\" (Version: 4.5.0) [\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e, \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e]. To make the model easier to read, XGBoost employed the SHapley Additive exPlanations (SHAP), an attractive extra interpretation technique. The output of any machine-learning model can be explained using the SHAP.We determined the SHapley Additive exPlanations (SHAP) values of the blue genes in two machine learning models, XGBoost and LightGBM, using the \"shapviz\" R package (Version: 0.9.5).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.7 Construction of protein-protein interaction (PPI) network and screening of hub genes\u003c/h2\u003e \u003cp\u003eUsing Relevance Scores larger than 2.5, 7600 genes linked to mitochondrial dysfunction were evaluated in GeneCards (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.genecards.org/\u003c/span\u003e\u003cspan address=\"https://www.genecards.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e)\u003c/span\u003e (Supplementary Table\u0026nbsp;1). The relevance of each of these genes to the phenotype is determined by looking at their Relevance Scores. The modular genes were imported into the STRING database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://cn.string-db.org/\u003c/span\u003e\u003cspan address=\"https://cn.string-db.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e)\u003c/span\u003e in order to determine the link between protein-protein interactions. After importing the relationship's TSV file into Cytoscape (Version: 3.8.1) software, two Cytoscape plug-ins\u0026mdash;Cytohubba (Version: 2.0.2) and MCODE (Version: 0.1)\u0026mdash;were used to choose the feature genes [\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e]. While MCODE was chosen based on the degree cutoff value\u0026thinsp;=\u0026thinsp;2 and maximum depth\u0026thinsp;=\u0026thinsp;100 Selection, Cytohubba was chosen using the EcCentricity method. The most described genes are those that arise from taking the intersection of the screened feature genes of MCODE and Cytohubba, the genes associated to mitochondrial function, and the machine learning screened the most .\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e2.8 Validation ladder of most characterized gene (MCG)\u003c/h2\u003e \u003cp\u003eUsing the R package \"pROC\" (Version:1.18.0), we mapped the diagnostic value of the most characterized gene. In order to create a validation ladder, we combined the two datasets, GSE1145 and GSE32453, and looked at the expression of the most characteristic gene (MCG).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e2.9 Establishment of a 28-day pathological cardiac hypertrophy model by Transverse Aortic Constriction (TAC)\u003c/h2\u003e \u003cp\u003e For each animal, the following 3 different researchers were involved: the first researcher performed the TAC procedure according to the randomization table. This researcher was the only one who knew the TAC group assignment. The second researcher was in charge of the mouse anesthesia procedure, and the third researcher was in charge of the surgical procedure. The Xiamen University Committee for the Keeping and Use of Animals approved the entire experiment, which followed the guidelines for the keeping and use of laboratory animals published by the National Institutes of Health of the United States of America and the Ministry of Health of the People's Republic of China. The aortic arch can be narrowed to simulate cardiac hypertrophy. The C57BL/6 wild-type mice utilized in this study were kept in the SPF-grade environment of the Xiamen University Animal Experiment Center, which was acquired from Beijing VitalRiver. The body weight was kept between 22 and 25 grams. The mice were raised in typical lighting circumstances (12 h/12 h dark cycle), with stable temperature ranges of 22\u0026ndash;25\u0026deg;C, 50% relative humidity, and unrestricted access to food and drink. Mice were weighed and given an intraperitoneal injection of sodium pentobarbital at a dosage of 50 mg/kg to induce anesthesia. The mice's tails and limbs were fastened, and they were positioned supine. The anterior thorax of the mice was hairless. The aortic arch was liberated by expanding the two broken ends of the mouse sternum along its edge into the second intercostal space under a body microscope. Then, the saline-impregnated 6\u0026thinsp;\u0026minus;\u0026thinsp;0 suture was fed through and into the posterior aspect of the aortic arch with a threader, and the 27G cushion needle was positioned parallel to the aortic arch, rapidly ligating it. Following that, the cushion needle was quickly removed in order to seal the thoracic cavity and any injuries. Iodophor was applied to the surgery site's skin to disinfect it. The mice were housed in an animal observation room with consistent humidity and temperature after they were awakened. Samples of animal tissue were taken 28 days following the models. With the exception of aortic arch ligation, the mice in the sham group underwent the same procedures as the mice in the TAC group. Animals were randomized after surviving the initial TAC, using a computer-based random order generator. Mice were put to sleep with sodium pentobarbital (50 mg/kg) four weeks following the operation, cervical subluxation method, and their hearts were removed for examination. The following parameters were assessed: heart weight to body weight (BW), heart weight to tibia length (TL), HW/TL (mg/mm), and HW/BW (mg/g). The number of mice in each group was six. Mice were included in the TAC success model if they did not die after 28 days and if they died during the 28-day period. or the 27G cushion needle was dislodged, they were not included in the statistics.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e2.10 Mouse heart ultrasound\u003c/h2\u003e \u003cp\u003eThe operator shaved the anterior chest of each mouse after administering 1.5% isoflurane to induce anesthesia. An ultrasound coupling agent was put on the chests of the mice and the ultrasound probe. The VisualSonics high-resolution Vevo 2100 system (VisualSonics, Toronto, Canada) was used to record and count the echocardiographic data from the mice.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e2.11 Hematoxylin-eosin (H\u0026amp;E)staining, Masson's Trichrome staining, Picro-Sirius Red Stain and Wheat Germ Agglutinin (WGA) staining\u003c/h2\u003e \u003cp\u003eEvery segment was situated within the papillary muscle's plane. H\u0026amp;E staining: We used the following procedure to stain the excised cardiac specimens with H\u0026amp;E. The stages of deparaffinization were as follows: We rinsed the sections with tap water, treated them with hematoxylin fractionation solution, and then rinsed them again with tap water. Xylene was used twice, for a total of 20 minutes each time; twice in 100% ethanol for 5 minutes and once in 75% ethanol for 5 minutes. 85% ethanol for five minutes, 95% ethanol for five minutes, and lastly dyeing the sections with eosin dye for five minutes to seal them with neutral gel. We observed and took pictures under a microscope.\u003c/p\u003e \u003cp\u003eMasson's trichrome staining procedure: paraffin slices were dewaxed in water, and then they were rinsed three times using tap and distilled water, respectively. This was immediately followed by the staining of cell nuclei for ten minutes using Regaud's hematoxylin stain and then another three times with tap and distilled water. Subsequently, the nuclei were dyed for eight minutes using Masson Lichun red acidic compound red solution. After differentiating for 4 minutes with a 1% phosphomolybdic acid aqueous solution, the cells were first rinsed with a 2% glacial acetic acid aqueous solution. They were then immediately stained for 5 minutes with either aniline blue or optic green solution, and they were then rinsed again with 0.2% glacial acetic acid aqueous solution. Finally, 95% alcohol, neutral gum sealing, xylene clear, and anhydrous alcohol.\u003c/p\u003e \u003cp\u003eSteps for Picro-Sirius Red Stain: First, dewax: five minutes for xylene Ⅰ, five minutes for xylene Ⅱ, and five minutes for xylene Ⅲ; then, anhydrous alcohol for one minute; then, 95% ethanol for one minute; and lastly, distilled water washing for five minutes. Subsequently, droplets of ferric hematoxylin staining solution are left for 5\u0026ndash;10 minutes. The surplus staining solution is then washed out with tap water for 5 minutes, and then drops of Sirius red staining solution are left for 15\u0026ndash;30 minutes. A mild rinsing was done with running water. For one minute, each of the following treatments: 75%, 95%, and anhydrous ethanol; three times, for one to two minutes each; and neutral gum sealing. Results of microscopic examination showed that muscle fibers were yellow, collagen fibers were red, and cell nuclei were tan. Among them, type III collagen fibers were green, while type I collagen fibers were strongly orange-yellow or brilliant red.\u003c/p\u003e \u003cp\u003eThe steps involved in WGA staining were as follows: paraffin sections were dewaxed to water, and then they were soaked in Eco-friendly dewaxing solutions I, II, and III for ten minutes each; they were then soaked in anhydrous ethanol Ⅰ for five minutes, anhydrous ethanol Ⅱ for five minutes, and anhydrous ethanol Ⅲ for five minutes; finally, they were rinsed with distilled water. Antigen repair: To repair against antigens, tissue pieces were placed in a microwave oven inside a repair cassette that was filled with EDTA antigen repair buffer (PH8.0). After 8 minutes of boiling at 55\u0026deg;C, switch off the fire and let it cool naturally for 7 minutes at 45\u0026deg;C. Don't forget to dry the slides. Slide in PBS (PH7.4) should be placed on a decolorizing shaker. Shake and wash three times, for five minutes each time. Using a histochemical pen, draw a circle around the tissue to keep the antibody from dripping off. Then, add the diluted WGA staining solution and let it to incubate for one hour at 37\u0026deg; in a thermostated room away from light. DAPI restaining of the nuclei: On a decolorizing shaker, place the slide in PBS (PH7.4) and wash three times for five minutes each. Add the DAPI staining solution, then let it sit at room temperature and keep it out of the light for ten hours. Tissue autofluorescence quenching procedure: three PBS (PH7.4) washes of the slides were performed on a decolorizing shaker for five minutes each. After adding the autofluorescence quencher B, the slides were rinsed for ten minutes under running water. sealing: Anti-fluorescence quenching sealer was used to seal the slides. Image acquisition: 488 channel positive is green, while DAPI channel nuclei are blue.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e2.12 Neonatal mouse primary cardiomyocytes isolation, culture and stimulation\u003c/h2\u003e \u003cp\u003eFollowing complete alcohol sterilization of the C57BL/6 suckling mice, the mice were pinched with the left hand to reveal their chest, and the ribs were sliced upward along the left bottom edge of the sternal raphe using a pair of ophthalmic straight scissors with the right hand. The suckling mouse's heart was extruded, and the left hand was slightly pushed up. Next, using ophthalmic forceps, the ventricular portion of the heart was cut off straight from the middle and placed into a Petri dish containing 4\u0026deg;C pre-cooled PBS. The fibrous tissue and blood clot in the Petri dish were then removed, and the heart tissue was evenly cut into pieces using curved ophthalmic scissors. The broken heart should be transferred to a 50 ml centrifuge tube. The PBS should be discarded. The tube should then be washed twice with PBS, collagenase type II should be added to submerge the heart, and the digestion process should be completed by shaking the heart for 5 minutes in a 37\u0026deg;C water bath. Use the 10% medium containing serum to complete the digestion. Finally, the heart should be shaken three times for 5 minutes. Following that, the entire centrifuge tube was centrifuged for five minutes at a rotational speed of 1000 rpm. Plates were arranged in Petri dishes following centrifugation. The supernatant medium was thrown out three hours after the cells were attached. After 48 or 72 hours of incubation, treatment ingredients might be added, depending on the needs of the experiment. The culture medium used for primary cardiomyocytes was DMEM high glucose supplemented with 10% bovine fetal serum and 1% penicillin-streptomycin. Six-well plates containing uniformly spaced cells were treated for 24 hours with 1 \u0026micro;M angiotensin II (A9525, Sigma, USA) in each well. The control group added complete medium without angiotensin II. We then gathered cells for additional tests.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e2.13 Extraction of sample RNA and RT-PCR\u003c/h2\u003e \u003cp\u003eOne milliliter of TRIzol was added to the mouse heart tissue and neonatal mouse primary cardiomyocytes, which were then crushed using a homogenizer, a cell scraper, and the heart tissue itself. After 10 minutes, 0.2 mL of chloroform was added, and the entire mixture was centrifuged for 15 minutes at 4\u0026deg;C at 12,000 rpm after being shaken for 30 seconds and allowed to stand at 22\u0026deg;C for 10 minutes. The material was separated into three layers by centrifugation, and the aqueous phase of the top layer was moved to a fresh tube. To precipitate the RNA, an equal volume of isopropanol was added and allowed to sit at room temperature for ten minutes. At 4\u0026deg;C, centrifugation was carried out for 10 minutes at 12,000 rpm. Centrifugation produced an appearance of the RNA precipitate. After removing the supernatant, 1 milliliter of 75\u0026deg;C ethanol was used to wash the RNA precipitate. Repeat the above twice, then centrifuge at 12,000 rpm for 5 minutes at 4\u0026deg;C. Discard the supernatant. After adding DEPC water to solubilize RNA and reverse RNA (LC480, USA), real-time PCR analysis was carried out. The sequences of primers are displayed in Supplementary Table\u0026nbsp;2.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e2.14 Western blot\u003c/h2\u003e \u003cp\u003eFollowing the application of RIPA lysis buffer, including protease and phosphatase inhibitors, to the cells and cardiac tissues, the protein stock solution was extracted, and the protein concentration was ascertained using a bicinchoninic acid (BCA) protein kit. We used an electrotransfer tank to move proteins to polyvinylidene difluoride (PVDF) membranes after separating them with 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). The PVDF membrane was sealed for one hour with 5% skim milk, and then it was incubated with the primary antibodies GAPDH (1:10,000, Proteintech Cat. No. 60004-1-Ig), ANP (1:1,000, Proteintech Cat. No. 27426-1-AP), and Myh7b (1:1,000, Cat. No. ab172967, Abcam) for the entire night at 4\u0026deg;C. We subsequently subjected them to an enhanced chemiluminescence (ECL) luminescence solution for additional examination.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e2.15 Tunel staining\u003c/h2\u003e \u003cp\u003eWe sequentially placed the slices in xylene I for 10 minutes, xylene II for 10 minutes, and xylene III for 10 minutes, followed by anhydrous ethanol I for 5 minutes, anhydrous ethanol II for 5 minutes, and anhydrous ethanol III for 5 minutes, after which they were washed with distilled water. To keep the antibody from leaking, lightly dry the cell slide and use a Liquid Blocker PAP Pen to make a circle in the middle of the cover glass where the cells are evenly spread out. Next, cover the tissue with a protease K working solution and place it in an incubator at 37\u0026deg;C for 22 minutes. Wash the slide with PBS (pH 7.4) in a decoloring shaker for 3 times, 5 minutes each time. The method for preparing the working solution of protease K involves a 1:1 ratio of PBS to stock solution. Dry the cell slide slightly, then add permeabilizing working solution to cover the tissue, incubate at room temperature for 20 min, and wash with PBS 3 times for 5 min each time (the permeabilizing working solution is 0.1% Triton). Configuration method, Triton stock solution: PBS\u0026thinsp;=\u0026thinsp;1:1000). Equilibrium at room temperature: After the climbing slice was slightly dried, buffer was dripped into the circle to cover the tissue, and the buffer was incubated at room temperature for 10 min. Tunnel reaction: Take the appropriate amount of TDT enzyme, dUTP, and buffer in the Tunnel kit according to the number of slices and tissue size and mix at a 1:5:50 ratio, and add to the circle to cover tissue. In a flat wet box, incubate at 37\u0026deg;C for 1 h. Be sure to keep the wet box moist by adding water. DAPI counterstain in nucleus: Wash three times with PBS (pH 7.4) in a decoloring shaker for 5 minutes each time. After removing PBS, a DAPI solution was dripped into the circle and incubated at room temperature for 10 min in the dark. Mount: Wash three times with PBS (pH 7.4) in a decoloring shaker, 5 minutes each time. After the slice is slightly dried, then cover it with an anti-fade mounting medium. Microscopy detection and collect images. DAPI glows blue by UV excitation wavelength.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e2.16 Targeted Drugs for Predicting HCM\u003c/h2\u003e \u003cp\u003eConnectivityMap (Cmap) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://clue.io/\u003c/span\u003e\u003cspan address=\"https://clue.io/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) counts the cellular gene expression of various small molecule interferences, such as certain chemical monomers or small molecule proteins, following stimulation and treatment of various human cells to create a database in which drug molecules are closely associated with gene expression[\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e]. Cmap mostly counts various Cmap primarily counts the many gene alterations brought on by various substances activating various cell types [\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e]. In order to further filter the medications, we first computed the standardized connectivity score (CS) by screening the differently expressed genes. Then, we calculated the score for each drug in relation to the differentially expressed genes [\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e]. We used the R programming language to view and evaluate the blue module differential genes after importing them into Cmap and using a computation to select the right medications.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003e2.17 Molecular docking for predicting drug-to-molecule action\u003c/h2\u003e \u003cp\u003eWe obtained the predicted drug molecule files from ChemSpider (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e\u003ca href=\"https://www.ncbi.nlm.nih.gov/geo/\" target=\"_blank\"\u003ewww.chemspider.com\u003c/a\u003e\u003c/span\u003e\u003cspan address=\"http://www.chemspider.com\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). We obtained protein structures of the most characterized genes from the Research Collaboratory for Structural Bioinformatics (RCSB) Protein Data Bank (PDB) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.rcsb.org/\u003c/span\u003e\u003cspan address=\"https://www.rcsb.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). Docking of drugs to protein structures was realized by AutoDockTools software(Version: 1.5.7). [\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e]. AutoDock uses a semi-flexible docking method that lets small molecules change their shape, and it judges the success of the docking by the binding free energy. We further optimized the protein to ensure chemical correctness and optimized the protein structure for docking. And we demonstrate this by drawing pictures of molecular docking using PyMOL software(Version:3.0).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003e2.18 Statistical analysis\u003c/h2\u003e \u003cp\u003eAll findings are presented as mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation (SD), and GraphPad (GraphPad Prism software, version 8.0.1) was used for analysis. We evaluated the differences in each sample using two-tailed t-tests or rank sum tests with unpaired samples. It was deemed significant when p\u0026thinsp;\u0026lt;\u0026thinsp;0.05. We conducted at least three replications of each experiment.\u003c/p\u003e \u003c/div\u003e"},{"header":"3 Results","content":"\u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Screening for different expression genes (DEGs) and screening for the most relevant modular genes\u003c/h2\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e1\u003c/span\u003e displays all of the article's experimental concepts. The GEO database gave us the transcriptome data and clinical information of GSE36961, GSE1145, and GSE32453. We split them into two groups, one for hypertrophic cardiomyopathy and the other for control. The RNA-seq data was cleaned and standardized by R package \"limma\" (Version:3.56.2) (Supplementary Fig.\u0026nbsp;1A, B). To identify differentially expressed genes (DEGs), we applied a rigorous filtering criterion: genes with an adjusted p-value (FDR)\u0026thinsp;\u0026lt;\u0026thinsp;0.05 and an absolute log2 fold change |log2FC| \u0026gt; 1 were considered significant. Using this approach, we identified 4,025 upregulated genes and 3,950 downregulated genes in the experimental ladder (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e2\u003c/span\u003eA, B). We used a Weighted Gene Co-expression Network Analysis (WGCNA) analysis on the DEGs expression profile matrix that we had just made. We split the DEG expression profile matrices into groups for the control group and the groups with hypertrophic cardiomyopathy (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e2\u003c/span\u003eC). By determining the ideal soft threshold, we can build a WGCNA that is more in line with scale-free. To enhance network robustness by minimizing spurious correlations and background noise, the adjacency matrix was transformed into a Topological Overlap Matrix (TOM) using a soft thresholding power of β\u0026thinsp;=\u0026thinsp;10. This parameter was selected as the minimum power value at which the scale-free topology fit index (R2) reached a plateau (R2\u0026thinsp;=\u0026thinsp;0.86), ensuring optimal network connectivity while preserving biological relevance. Subsequently, the topological dissimilarity measure (1\u0026thinsp;\u0026minus;\u0026thinsp;TOM) was calculated to facilitate downstream module detection through hierarchical clustering. Selecting β\u0026thinsp;=\u0026thinsp;10 and R\u0026sup2;=0.86 (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e2\u003c/span\u003eD, E) as our soft threshold, we converted the adjacency into a topological overlap matrix (TOM). We converted the matrix into 14 gene modules using the dynamic shear approach (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e2\u003c/span\u003eF). In the end, 36 black module genes, 200 blue module genes, 22 brown module genes, 60 green module genes, 34 green-yellow module genes, 15 gray-60 module genes, 14 light cyan module genes, 14 light green module genes, 5 light yellow module genes, 4 magenta module genes, 12 midnight blue module genes, 13 pink module genes, and 15 yellow module genes were taken out (Supplementary Table\u0026nbsp;3). This was achieved by setting thresholds of 0.8 for membership (MM), 0.1 for gene significance, and 0.1 for weight. For cluster analysis, we put the module genes into groups so that we could see what the link was between modules and clinical data (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e2\u003c/span\u003eG). In order to look into the relationship between modules and clinical symptoms, we examined the link between Module Eigengene E (ME) and clinical aspects. We found the correlation coefficient (cc) between different modules and clinical characteristics to see how strong the link was (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e2\u003c/span\u003eH). Another meaningless module, the blue one, had the weakest correlation of the bunch. It had a negative correlation with HCM (cc = -0.83, p\u0026thinsp;=\u0026thinsp;8.4e-39) and a positive correlation with Control (cc\u0026thinsp;=\u0026thinsp;0.83). The Grey module held little significance. It is important to note that there may be a relationship between sex and the magenta and green-yellow module genes. It was also interesting to see how the gene significance of the modules related to their membership. The blue module had the strongest link (r\u0026thinsp;=\u0026thinsp;0.78, p\u0026thinsp;=\u0026thinsp;0.0e\u0026thinsp;+\u0026thinsp;0) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e2\u003c/span\u003eI). Supplementary Fig.\u0026nbsp;1C-N displays more module-specific correlation graphs.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec23\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Functional enrichment analysis\u003c/h2\u003e \u003cp\u003eWe performed GO, KEGG, and GSEA functional enrichment analyses on the genes in the blue module. This gave us additional insight into the function of the blue module in HCM. To find out more about the blue module genes, we first performed GO enrichment scores by biological process (BP), cellular component (CC), and molecular function (MF).Actin filament organization, positive regulation of defense response, leukocyte cell- cell adhesion, regulation of inflammatory response, myeloid leukocyte activation, myeloid cell differentiation, response to bacterial molecules, response to lipopolysaccharide, lymphocyte proliferation, and myeloid cell activation involved in immune response were the main biological process-level enrichment zones of the blue module; Focused adhesion, cell-substrate junction, membrane raft, membrane microdomain, secretory granule lumen, cytoplasmic vesicle lumen, vesicle lumen, primary lysosome, azurophil granule, and nuclear periphery are among the cellular components that the blue module is primarily enriched with. On the molecular function side, the blue module focuses on actin binding, organic anion transmembrane transporter activity, actin filament binding, calcium-dependent protein binding, non-membrane spanning protein tyrosine kinase activity, phosphatidylinositol 3-kinase binding, RAGE receptor binding, Toll-like receptor binding, and long-chain fatty acid binding (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). Additionally, we performed a KEGG enrichment analysis to determine the function of the pathway it resides in. Salmonella infection, apoptosis, tight junction, pathogenic Escherichia coli infection, gap junction, Fc gamma Rmediated phagocytosis, bacterial invasion of epithelial cells, pertussis, complement and coagulation cascades, and the Hippo signaling pathway were the findings that showed the blue module was primarily in the phagosome. Yersinia infection, tuberculosis, prion disease, NF-kappa B signaling route, hippo signaling pathwaymultiple species, mineral absorption, motor proteins, Shigellosis, and amyotrophic lateral sclerosis (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e3\u003c/span\u003eB, C). The Toll-Like Receptor Signaling Pathway, GnRH Signaling Pathway, Nod-Like Receptor Signaling Pathway, Natural Killer Cell-Mediated Cytotoxicity, B Cell Receptor Signaling Pathway, JAK/STAT Signaling Pathway, and Apoptosis were the main focus of the GSEA results, according to our GSEA enrichment analysis of the samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e3\u003c/span\u003eD).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003e3.3 The immune microenvironment of HCM\u003c/h2\u003e \u003cp\u003eWhen we found a strong link between the blue module's gene enrichment and the immune system, we looked into the immunological microenvironment of HCM. In that order, we used xCell, ESTIMATE, EPIC, Microenvironment Cell Populations-counter (MCP-counter), CIBERSORT, and ssGSEA to study the immune microenvironment of HCM. In the xCell results, the HCM group of iDC, Adc, Basophils, MicroenvironmentScore, ImmuneScore, Monocytes, cDC, Macrophages M1, Preadipocytes, NKT, DC, Macrophages, and mv Endothelial cells, B-cells, MEP, GMP, pro B-cells, CMP, plasma cells, CLP, hepatocytes, memory B-cells, neutrophils, CD4\u0026thinsp;+\u0026thinsp;T-cells, and astrocytes were less enriched compared to the control group, but neurons, CD8\u0026thinsp;+\u0026thinsp;naive T-cells, myocytes, CD8\u0026thinsp;+\u0026thinsp;Tcm cells, platelets, CD8\u0026thinsp;+\u0026thinsp;T-cells, chondrocytes, pericytes, HSC, osteoblasts, and smooth muscle were enriched to an elevated level (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e4\u003c/span\u003eA-C). The HCM group's scores were lower than the control group's for ImmuneScore, StromalScore, and ESTIMATES (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e4\u003c/span\u003eD). We used EPIC to look more closely at the enrichment of seven common immune-associated cells. The HCM group had lower levels of endothelial and macrophage cells, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e4\u003c/span\u003eE. Next, we used MCPcounter to measure the quantity of ten immune-related cells. Monocytic lineage, myeloid dendritic cells, neutrophils, and fibroblasts were reduced in abundance in the HCM group compared to the control group (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e4\u003c/span\u003eF). Based on the CIBERSORT data, there were fewer monocytes in the HCM group compared to the Control group, but more T cells CD8, NK cells at rest, and macrophages M2 (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e4\u003c/span\u003eG). We employed ssGSEA to look into the samples' level of immune infiltration in more detail. HCM group of APC co-stimulation, Myeloid-derived suppressor cell, regulatory T cell, Macrophage, Central memory CD8 T cell, mast cell, activated CD4 T cell, effector memory CD8 T cell, neutrophil, gamma delta T cell, parainflammation, activated dendritic cell, T cell co-inhibition, Eosinophil, CCR, Check-point, Memory B cell, Type 1 T helper cell, Central memory CD4 T cell, Immature B cell, T follicular helper cell, Activated B cell, HLA, Immature dendritic cell, T cell co-stimulation, Natural killer cell, Monocytes, type 2 T helper cells, CD56 bright natural killer cells, type 17 T helper cells, plasmacytoid dendritic cells, and CD56 dim natural killer cells have low abundance. Effector memory CD4 T cells have higher abundance (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e4\u003c/span\u003eH-K).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec25\" class=\"Section2\"\u003e \u003ch2\u003e\u003cb\u003e3.4 Machine Learning and Protein Levels Combined to Co-Screen the most characteristic gene (MCG)\u003c/b\u003e\u003c/h2\u003e \u003cp\u003eWe got the relevant expression profile matrices of the blue module genes and used machine learning to get the characterized genes. We conducted this process for each machine learning model in order to identify the genes with the highest level of characterization. We achieve sparsity in the Lasso Regression model by incorporating an L1-paradigm penalty term, which eliminates additional features. Through binomial deviation and rainbow plots (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e5\u003c/span\u003eA, B), we found that the best model had five coefficients that were not zero. It also looked at five genes at the same time. The Random Forest technique typically uses the IncMSE and IncNodePurity screens to assess gene characterization, but the IncNodePurity screen holds greater significance. Based on IncNodePurity, we evaluated 15 described genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e5\u003c/span\u003eC, D).\u003c/p\u003e \u003cp\u003eAdditionally, we employed a decision tree model to screen feature genes. Some of the genes that the model looked at were MT1M, CEBPD, ZFP36, TUBA3E, and CDC42EP4 (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e5\u003c/span\u003eE). CDC42EP4 was the switch between the control and HCM groups. We also chose to do 10-fold cross-validation 10 times, use SVM-REF to find the feature genes, plot the histograms, and do K-fold cross-validation 10 times using the \"random\" repeated cross-validation method with SIZE\u0026thinsp;=\u0026thinsp;1:10 (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e5\u003c/span\u003eF). Next, we used root mean square error (RMSE) to assess the regression model's accuracy. It was the CDC42EP4, CEBPD, CSRNP1, FCN3, MT1M, SERPINA3, TUBA3C, TUBA3D, TUBA3E, and ZFP36 group of genes that worked best. We identified them at the lowest RMSE (RMSE\u0026thinsp;=\u0026thinsp;6) point (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e5\u003c/span\u003eG). Next, we used the \"shapviz\" R package (Version: 0.9.5) to find the SHapley Additive exPlanations (SHAP) values of the blue genes in XGBoost and LightGBM, two machine learning models (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e5\u003c/span\u003eH\u0026ndash;J and Supplementary Fig.\u0026nbsp;2A, B). Additionally, we determined each defined gene's variable relevance and eliminated the 20 genes with the highest scores (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e5\u003c/span\u003eK, L). We also used SHAP in the XGBoost model to look at the connections between feature genes (Supplementary Fig.\u0026nbsp;2C). Subsequently, we performed additional protein-level screening of the feature genes. We loaded the blue module genes into the STRING database using Cytoscape (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e5\u003c/span\u003eM). Using two plugins, Cytohubba and MCODE, respectively, we screened the described genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e5\u003c/span\u003eN, O). Ultimately, the most well-characterized mitochondrial gene, CEBPD (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e5\u003c/span\u003eP), was obtained by crossing the genes screened using these seven techniques with genes related to mitochondrial dysfunction. Additionally, in two machine learning models, XGBoost and LightGBM, respectively, we found the SHAP dependence of CEBPD (Supplementary Fig.\u0026nbsp;2D, E).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec26\" class=\"Section2\"\u003e \u003ch2\u003e3.5 Validation ladders of the most characteristic gene\u003c/h2\u003e \u003cp\u003eAs a validation ladders dataset, we combined the HCM datasets GSE32453, GSE35229, and GSE1145 (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e6\u003c/span\u003eA). We looked at CEBPD expression using the validation ladder's transcript levels. We validated the experimental ladder results by finding lower CEBPD expression in the HCM group (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e6\u003c/span\u003eB). We then used ROC curves in the validation ladder and experimental ladder, respectively, to assess the diagnostic utility of the most described genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e6\u003c/span\u003eC, D).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec27\" class=\"Section2\"\u003e \u003ch2\u003e3.6 Alterations to the HCM's overall shape and functional phenotype\u003c/h2\u003e \u003cp\u003eAfter 28 days of modeling, heart tissues were extracted from mice. When comparing the TAC group to the sham group, the Left Ventricular Shortening Fraction (LVFS) and Left Ventricular Ejection Fraction (LVEF) were both lower (Fig.\u0026nbsp;\u003cspan refid=\"Fig14\" class=\"InternalRef\"\u003e7\u003c/span\u003eA, B). Next, anatomical features associated with the mouse cardiac hypertrophy model were examined. First, the morphology of the heart showed that TAC caused the heart to enlarge, unlike Sham (Fig.\u0026nbsp;\u003cspan refid=\"Fig14\" class=\"InternalRef\"\u003e7\u003c/span\u003eC). In the meanwhile, we took measurements of the mice's tibia length, heart weight, and body weight. Then, in order to determine cardiac hypertrophy, we calculated the ratios of heart weight to body weight (BW) and heart weight to tibia length (TL). The HW/TL (mg/mm) and HW/BW (mg/g) ratios were considerably greater in the TAC group (Fig.\u0026nbsp;\u003cspan refid=\"Fig14\" class=\"InternalRef\"\u003e7\u003c/span\u003eD). Tissue slice staining with WGA and H\u0026amp;E further confirmed the TAC-induced cardiac hypertrophy. Also, tissue slices stained with Picro-Sirius Red Stain and Masson showed that the ventricles had changed after the heart got bigger. According to our research, the TAC group exhibited higher volumes of cardiomyocytes and the fibrotic layer (Fig.\u0026nbsp;\u003cspan refid=\"Fig14\" class=\"InternalRef\"\u003e7\u003c/span\u003eE, F). Next, an experimental investigation was conducted on the expression of left ventricular genes associated with β-myosin heavy chain (β-MHC) and atrial natriuretic peptide (ANP). To see what the RT-PCR and Western blot tests showed, TAC greatly increased the levels of mRNA and protein expression of ANP and β-MHC (Fig.\u0026nbsp;\u003cspan refid=\"Fig14\" class=\"InternalRef\"\u003e7\u003c/span\u003eG, I). Another thing that was found was that when angiotensin II was added to newborn mouse primary cardiomyocytes, the levels of ANP and β-MHC went up, which was in line with the TAC data (Fig.\u0026nbsp;\u003cspan refid=\"Fig14\" class=\"InternalRef\"\u003e7\u003c/span\u003eH, J).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec28\" class=\"Section2\"\u003e \u003ch2\u003e3.7 Expression of MCG in vivo and in vitro\u003c/h2\u003e \u003cp\u003eThe TAC group had more cells that had already died because the blue module genes were linked to a higher risk of apoptosis (Fig.\u0026nbsp;\u003cspan refid=\"Fig16\" class=\"InternalRef\"\u003e8\u003c/span\u003eA, B). We checked the mRNA level of CEBPD in heart tissues to make sure that MCG was expressed in the animal model. We found that the TAC group had decreased CEBPD expression (Fig.\u0026nbsp;\u003cspan refid=\"Fig16\" class=\"InternalRef\"\u003e8\u003c/span\u003eC). Moreover, we validated this using in vitro experiments. Ang II was used to stimulate neonatal mouse primary cardiomyocytes. The levels of CEBPD mRNA were lowered in neonatal mouse primary cardiomyocytes that were stimulated with Ang II, just like they were in the animal model (Fig.\u0026nbsp;\u003cspan refid=\"Fig16\" class=\"InternalRef\"\u003e8\u003c/span\u003eD). In the end, we used western blotting to find CEBPD expression in animal tissues. The expression matched mRNA at lower protein levels (Fig.\u0026nbsp;\u003cspan refid=\"Fig16\" class=\"InternalRef\"\u003e8\u003c/span\u003eE, F).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec29\" class=\"Section2\"\u003e \u003ch2\u003e3.8 Exploring Targeted Drugs for HCM\u003c/h2\u003e \u003cp\u003eWe used the Cmap website to import the blue module genes that were filtered by WGCNA based on differential expression, and we used this database to search for particular target medications. We looked at the 10 compounds with the highest scores based on their normalized connectivity score (CS) and false discovery rate (FDR (nlog10)). They were Resiquimod, Tazarotene, Rimexolone, Medrysone, Adarotene, Abt-751, Dexketoprofen, Ciclesonide, Tolmetin, and Aripiprazole (Fig.\u0026nbsp;\u003cspan refid=\"Fig18\" class=\"InternalRef\"\u003e9\u003c/span\u003eA). We might use these medications to treat HCM. Next, we used OpenBabel software to transform the 10 compounds' chemical structures into 3D structures after downloading them from ChemSpider. Next, the RCSB website was used to get the CEBPD protein molecules. These were then worked on with PYMOL software and docked with AUTODOCK software (Fig.\u0026nbsp;\u003cspan refid=\"Fig18\" class=\"InternalRef\"\u003e9\u003c/span\u003eB-K). We discovered that all of the aforementioned chemicals have some affinity for CEBPD. Among these docked compounds, Abt-751 had the strongest affinity for CEBPD (Fig.\u0026nbsp;\u003cspan refid=\"Fig18\" class=\"InternalRef\"\u003e9\u003c/span\u003eL).\u003c/p\u003e \u003c/div\u003e"},{"header":"4 Discussion","content":"\u003cp\u003eLarger myocardial enlargement in hypertrophic cardiomyopathy blocks left ventricular outflow, which lowers the heart's energy supply and raises the risk of sudden death. We still don't fully grasp the course of treatment or the targets for hypertrophic cardiomyopathy, despite a great deal of research [\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e]. In order to find the appropriate modules for our investigation, we first collected DEGs from the patients and performed differential analysis. We conducted gene enrichment analysis using the genes from the blue module. Based on the gene enrichment analysis results, which pointed to a possible close link between immune cells and HCM development, we did an immune infiltration study. We then screened the most characteristic gene (CEBPD) using six machine learning filtering results, mitochondrial dysfunction-associated genes, and the outcomes of two Cytoscape software plug-ins. In line with the findings of the experimental ladder, we also examined the expression of CEBPD using the validation ladder. After that, we made HCM models in animals and cells and looked at the levels of HCM mRNA expression in living things and in cells. By forecasting and screening the blue module genes, we were able to identify ten medications that may have an impact on the development of HCM. In the end, we downloaded the CEBPD protein molecule and ten expected targeted medications to do molecular docking. We found that Abt-751, potentially a predicted targeted drug, exhibited the highest binding affinity with CEBPD.\u003c/p\u003e \u003cp\u003eOur article summarizes the following 4 points of finding:\u003c/p\u003e\u003cp\u003e1.CEBPD\u0026rsquo;s Role in Mitochondrial Dysfunction and Immune Crosstalk: Our multi-omics approach identified CEBPD as a central regulator linking mitochondrial dysfunction and immune dysregulation in HCM. Mechanistically, CEBPD downregulation may impair mitochondrial fusion-fission dynamics (supported by reduced oxygen consumption rates in vitro) and amplify pro-inflammatory responses via TLR-4/NF-κB signaling pathway [\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e]. This dual role aligns with recent studies showing CEBPD\u0026rsquo;s involvement in metabolic stress adaptation and macrophage polarization. The SHAP dependency plots (Supplementary Fig.\u0026nbsp;2D-E) further highlight CEBPD\u0026rsquo;s nonlinear interactions with inflammatory markers (e.g., IL-1β, COX-2), suggesting its potential as a therapeutic node.\u003c/p\u003e \u003cp\u003e2.Machine Learning Consensus Strategy: The integration of six machine learning algorithms (LASSO, SVM-RFE, RF, XGBoost, LightGBM, DecisionTree) ensured robust feature gene selection. For instance, LASSO prioritized sparsity (5 genes, RMSE\u0026thinsp;=\u0026thinsp;6; Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e5\u003c/span\u003eG), while XGBoost/LightGBM captured non-linear relationships through SHAP values (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e5\u003c/span\u003eH-J). Cross-validation across 10,000 permutations revealed that CEBPD consistently ranked in the top 0.5% of feature importance scores, underscoring its biological relevance.\u003c/p\u003e\u003cp\u003e3.Immune Microenvironment Insights: Immune infiltration analyses (xCell, CIBERSORT, ssGSEA) demonstrated significant depletion of monocytes and CD8\u0026thinsp;+\u0026thinsp;T cells in HCM (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e4\u003c/span\u003eG-K), coupled with elevated M2 macrophage polarization. These findings correlate with the blue module\u0026rsquo;s enrichment in \"myeloid leukocyte activation\" (GO:0043312; p\u0026thinsp;=\u0026thinsp;3.2e-10) and suggest that CEBPD may mediate immune-metabolic crosstalk via chemokine networks (e.g., CXCR4/SDF1 axis).\u003c/p\u003e \u003cp\u003e4.Therapeutic Implications: Molecular docking identified Abt-751 as the top candidate targeting CEBPD (binding energy: \u0026minus;9.8 kcal/mol; Fig.\u0026nbsp;\u003cspan refid=\"Fig18\" class=\"InternalRef\"\u003e9\u003c/span\u003eL). This aligns with Abt-751\u0026rsquo;s known inhibition of tubulin polymerization, a process implicated in mitochondrial trafficking. However, functional validation of Abt-751 in HCM models remains to be explored.\u003c/p\u003e \u003cp\u003eThe C/EBP family includes the adaptable transcription factor CCAAT/enhancer-binding protein delta (CEBPD), which is divided into three sections: the basic DNA-binding region, the C-terminal leucine-zipper domain, and the N-terminal transactivating region[\u003cspan additionalcitationids=\"CR64\" citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e]. This transcription factor is essential for controlling the expression of genes linked to inflammatory and immunological responses. According to Spek CA et al., CEBPD increased the inflammatory responses of macrophages; however, CEBPD-knockout macrophages were unable to identify the pro-inflammatory transcriptional pathway that is dependent on CEBPD [\u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e].That's why CEBPD controls IL-1β or collagen-induced cyclooxygenase-2 (COX-2) to work as a pro-inflammatory transcription factor and raise the levels of pro-inflammatory mediators. Furthermore, CEBPD triggers the expression of TLR-4 and the ensuing signaling [\u003cspan additionalcitationids=\"CR68 CR69 CR70 CR71\" citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e72\u003c/span\u003e]. In tumor immunity, CEBPD is able to enhance mRNA and protein expression via MMP2 in uroepithelial carcinomas, thereby increasing tumor invasiveness[\u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e73\u003c/span\u003e]. Furthermore, it has been discovered that the CEBPD-induced autophagy route is how metformin induces apoptosis in hepatocellular carcinoma cells [\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e]. In the cardiovascular area, Wang Q et al. found that vascular smooth muscle cell inflammation is regulated by a hierarchical and cooperative BRD4/CEBPD cooperation[\u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e75\u003c/span\u003e]. Chi JY's study showed that the fibroblast CEBPD/SDF4 axis responds to angiogenesis generated by chemotherapy via CXCR4 [\u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e76\u003c/span\u003e].Recent studies have focused on its connection to mitochondria, specifically on how it affects energy consumption and mitochondrial function. Genes involved in mitochondrial biogenesis have been linked to CEBPD regulation. It may change how important factors that encourage the development of new mitochondria are expressed, which would have an impact on the energy metabolism of the entire cell. Chan TC et al. discovered that CEBPD enhanced glucose uptake and lactate generation by upregulating SLC2A1 and HK2, resulting in mitochondrial fission, an elevated extracellular acidification rate, and a reduced oxygen consumption rate to support cellular proliferation [\u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e77\u003c/span\u003e]. Wang WJ et al. discovered that the transcription factor C/EBP δ (CEBPD) responds to active STAT3 (pSTAT3) and facilitates the transcriptional activation of MCL1 following leptin therapy. MCL1 facilitates leptin-induced mitochondrial fusion and correlates with GBC cell viability [\u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e78\u003c/span\u003e]. Banerjee et al. discovered a new study illustrating a unique function of C/EBPδ in safeguarding against both baseline and ionizing radiation-induced oxidative stress and mitochondrial dysfunction, hence enhancing post-radiation survival [\u003cspan citationid=\"CR79\" class=\"CitationRef\"\u003e79\u003c/span\u003e]. Recent study indicates that CEBPD may be implicated in mitochondrial dysfunction. Dysregulation of CEBPD may increase the pathophysiology of sickness, potentially impacting mitochondrial function. CEBPD is recognized for its regulation of genes associated with inflammatory reactions, potentially affecting mitochondrial function. Inflammatory cytokines may elicit mitochondrial alterations, and CEBPD can regulate these effects [\u003cspan citationid=\"CR80\" class=\"CitationRef\"\u003e80\u003c/span\u003e, \u003cspan citationid=\"CR81\" class=\"CitationRef\"\u003e81\u003c/span\u003e]. With the exception of preliminary research, the precise function of CEBPD in HCM is yet unknown. As a result, CEBPD is intriguing for the advancement of HCM diagnostic therapy and may be studied further in the future.\u003c/p\u003e \u003cp\u003eComprising endothelial cells, immune cells, and diverse fibroblasts, the heart is a complicated multicellular organ. The significance of immune cells in the molecular processes of cardiovascular disease has drawn more attention, especially among these non-cardiomyocytes [\u003cspan citationid=\"CR82\" class=\"CitationRef\"\u003e82\u003c/span\u003e, \u003cspan citationid=\"CR83\" class=\"CitationRef\"\u003e83\u003c/span\u003e]. An increasing body of research has demonstrated the role immune cells play in physiological processes such as heart formation, heart homeostasis maintenance, and heart aging [\u003cspan citationid=\"CR84\" class=\"CitationRef\"\u003e84\u003c/span\u003e, \u003cspan citationid=\"CR85\" class=\"CitationRef\"\u003e85\u003c/span\u003e]. The pathogenesis of heart remodeling and hypertrophy depends on immune cells [\u003cspan citationid=\"CR86\" class=\"CitationRef\"\u003e86\u003c/span\u003e]. Our blue module gene enrichment results showed a strong correlation between blue module gene s and immunity. By integrating many immune enrichment tests, we were able to examine the immune cell infiltration of the patients. There is a strong link between HCM and monocytic lineage, endothelial cells, natural killer T cells, memory B cells, immature B cells, ImmuneScore, and CD8\u0026thinsp;+\u0026thinsp;T cells. However, more research is necessary to pinpoint the exact mechanisms involved.\u003c/p\u003e \u003cp\u003eFinally, we found the ten most likely drugs that had targets by using targeted drug prediction for the blue module genes that were screened from the data. To forecast the binding ability, we molecularly docked each of these 10 potential targeted medicines with CEBPD. Abt-751 had the strongest binding ability, which suggests it could be used as a new HCM-targeting drug. We did not, however, confirm this with specific medications, which would have provided persuasive evidence for more research. Animals can initially receive these drugs intraperitoneally or by gavage to examine their impact on HCM. In vitro research can be done concurrently to look at the medications' actual mechanism of action. After that, we want to gather samples of human HCM to look at the impact of CEBPD on HCM development. Future studies on CEBPD can be conducted in humans and animals, respectively, to learn more about their distinct processes and the genetically related antagonists. This study could help discover how to prevent and treat CEBPD, as well as how to help doctors manage patients and predict their recovery.\u003c/p\u003e \u003cp\u003eHowever, our current study still has many limitations:Dataset Heterogeneity: While integrating GSE36961, GSE1145, and GSE32453 improved statistical power, batch effects from different platforms (GPL15389 vs. GPL570) may introduce bias. Our normalization mitigated but did not fully eliminate platform-specific variances. The number of Samples in our validation dataset is relatively small, and we need to increase the number of patients to validate the depth sufficiently; Experimental Model Constraints: The TAC-induced mouse model and Ang II-treated cardiomyocytes primarily reflect pressure-overload hypertrophy, which may not fully recapitulate genetic HCM pathophysiology. Future studies should validate CEBPD in human HCM tissues and MYH7-mutant models; Translational Gaps: Although Abt-751 showed strong in silico binding to CEBPD, its efficacy and safety in vivo remain unverified. Additionally, the immune infiltration results (e.g., M2 macrophage dominance) were derived from bulk RNA-seq data, which lacks single-cell resolution to delineate tissue-specific immune subsets; Mechanistic Depth: While we identified CEBPD\u0026rsquo;s association with mitochondrial dysfunction, the precise molecular pathways (e.g., mitophagy) require further CRISPR-based functional assays.\u003c/p\u003e \u003cp\u003eThese limitations highlight opportunities for future work, including single-cell omics in human HCM samples and preclinical trials of Abt-751. Nevertheless, our integrative approach provides a foundational framework for understanding CEBPD\u0026rsquo;s role in HCM and its potential as a therapeutic target.\u003c/p\u003e \u003cp\u003eIn summary, we have screened genes associated to mitochondrial dysfunction from transcriptome level to protein level for the first time in HCM by combining bioinformatics, multiple machine learning, and protein network level screening techniques. Furthermore, in order to fully explore the potential immunological targets of HCM, we integrated a number of the most widely used immune infiltration assays to examine its immune microenvironment. In the end, we made models of cells and animals to test our ideas, predicted possible medicines, and used molecular docking to give us a clue in our search for medicines that target HCM.\u003c/p\u003e"},{"header":"5 Conclusion","content":"\u003cp\u003eWe used PPI networks and a variety of machine learning techniques to screen for the novel target marker (CEBPD) of HCM. We came to the conclusion that CEBPD may function in HCM through the monocytic lineage, endothelial cells, natural killer T cells, memory B cells, immature B cells, and CD8\u0026thinsp;+\u0026thinsp;T cells when combined with immuno-infiltration analysis of the samples. The validation ladder's CEBPD expression matched that of the experimental ladder. In vitro and in vivo, CEBPD mRNA levels were also lowered. We also thought that the following ten drugs might be HCM targets: Aripiprazole, Tazarotene, Rimexolone, Medrysone, Adarotene, Abt-751, Dexketoprofen, Ciclesonide, Tolmetin, and Resiquimod. Finally, we found that Abt-751 has the highest binding affinity with CEBPD, indicating its potential as a targeted drug.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe paper has been reviewed and approved by all authors. JC, DX, and YL came up with the original concept for this research. ZW and RL oversaw the procedure and evaluated the final article. The data analysis and gathering were handled by ZC. The table and manuscript were created by XC, RH, and LL. Graphs with LX and SJ plotted. ZW examined the info. The study was planned and the text was revised by FL, JW, and ZS.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was supported by the Natural Science Foundation of China (Grant Nos.82070291). The funder had no role in the decision to publish or preparation of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank Feng-chun Lu, Jia-mao Wang and Zhong-gui Shan for technology supporting and composition instruction.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSupplementary Materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe full contents of the supplement are available online.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study and included experimental procedures were approved by the institutional animal care and use committee of Xiamen university (approval no. XMULAC20200137). All animal housing and experiments were conducted in strict accordance with the institutional guidelines for care and use of laboratory animals. Patient data from the GEO database were used in this study.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets supporting our findings are presented in the article. GSE36961, GSE1145 and GSE32453 were downloaded from the GEO database (https://www.ncbi.nlm.nih.gov/geo/). All analyzed data are included in this published article. The original data are available.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research utilized published studies and consortia that have made their summary statistics publicly available. All original studies included in this research have obtained approval from their respective ethical review boards, and participants have provided informed consent. It is important to note that no individual-level data was utilized in this study. As a result, no new ethical review board approval was necessary for this research.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThere are no conficts of interest to declare.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDisclosure statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNo potential conflict of interest was reported by the authors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical trial number\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor details\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003csup\u003e1\u003c/sup\u003eThe First Affiliated Hospital of Xiamen University, School of Medicine Xiamen University, NO.55, Zhenhai Road, Siming District, Xiamen, Fujian ,361003, China\u003c/p\u003e\n\u003cp\u003e\u003csup\u003e2\u003c/sup\u003e Department of General Surgery, Fujian Medical University Union Hospital, No. 29 Xinquan Road, Fuzhou, 350001, China.\u003c/p\u003e\n\u003cp\u003e\u003csup\u003e3\u003c/sup\u003e Department of Cardiac Surgery, Xiangan Hospital of Xiamen University, School of Medicine, Xiamen University, Xiamen, Fujian ,361100, China.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eNakamura, M. \u0026amp; Sadoshima, J. Mechanisms of physiological and pathological cardiac hypertrophy. \u003cem\u003eNat. Reviews Cardiol.\u003c/em\u003e \u003cb\u003e15\u003c/b\u003e (7), 387\u0026ndash;407 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcKenna, W. J., Maron, B. J. \u0026amp; Thiene, G. Classification, Epidemiology, and Global Burden of Cardiomyopathies. \u003cem\u003eCircul. Res.\u003c/em\u003e \u003cb\u003e121\u003c/b\u003e (7), 722\u0026ndash;730 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShimizu, I. \u0026amp; Minamino, T. Physiological and pathological cardiac hypertrophy. \u003cem\u003eJ. Mol. Cell. Cardiol.\u003c/em\u003e \u003cb\u003e97\u003c/b\u003e, 245\u0026ndash;262 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLyon, R. C., Zanella, F., Omens, J. H. \u0026amp; Sheikh, F. Mechanotransduction in Cardiac Hypertrophy and Failure. \u003cem\u003eCircul. Res.\u003c/em\u003e \u003cb\u003e116\u003c/b\u003e (8), 1462\u0026ndash;1476 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWeber, K. T., Clark, W. A., Janicki, J. S. \u0026amp; Shroff, S. G. Physiologic versus pathologic hypertrophy and the pressure-overloaded myocardium. \u003cem\u003eJ. Cardiovasc. Pharmacol.\u003c/em\u003e \u003cb\u003e10\u003c/b\u003e (Suppl 6), S37\u0026ndash;50 (1987).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChien, K. R., Knowlton, K. U., Zhu, H. \u0026amp; Chien, S. Regulation of cardiac gene expression during myocardial growth and hypertrophy: molecular studies of an adaptive physiologic response. \u003cem\u003eFASEB journal: official publication Federation Am. Soc. Experimental Biology\u003c/em\u003e. \u003cb\u003e5\u003c/b\u003e (15), 3037\u0026ndash;3046 (1991).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYamamoto, S., Kita, S., Iyoda, T., Yamada, T. \u0026amp; Iwamoto, T. New Molecular Mechanisms for Cardiovascular Disease: Cardiac Hypertrophy and Cell-Volume Regulation. \u003cem\u003eJ. Pharmacol. Sci.\u003c/em\u003e \u003cb\u003e116\u003c/b\u003e (4), 343\u0026ndash;349 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBernardo, B. C., Weeks, K. L., Pretorius, L. \u0026amp; McMullen, J. R. Molecular distinction between physiological and pathological cardiac hypertrophy: Experimental findings and therapeutic strategies. \u003cem\u003ePharmacol. Ther.\u003c/em\u003e \u003cb\u003e128\u003c/b\u003e (1), 191\u0026ndash;227 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAshrafian, H., Redwood, C., Blair, E. \u0026amp; Watkins, H. Hypertrophic cardiomyopathy: a paradigm for myocardial energy depletion. \u003cem\u003eTrends Genet.\u003c/em\u003e \u003cb\u003e19\u003c/b\u003e (5), 263\u0026ndash;268 (2003).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eElliott, P. \u0026amp; McKenna, W. J. Hypertrophic cardiomyopathy. \u003cem\u003eLancet\u003c/em\u003e \u003cb\u003e363\u003c/b\u003e (9424), 1881\u0026ndash;1891 (2004).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOkajima, Y., Tanabe, Y., Takayanagi, M. \u0026amp; Aotsuka, H. A follow up study of myocardial involvement in patients with mitochondrial encephalomyopathy, lactic acidosis, and stroke-like episodes (MELAS). \u003cem\u003eHeart\u003c/em\u003e \u003cb\u003e80\u003c/b\u003e (3), 292\u0026ndash;295 (1998).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMudhar, H. S., Wagner, B. E. \u0026amp; Suvarna, S. K. Electron microscopy of myocardial tissue. A nine year review. \u003cem\u003eJ. Clin. Pathol.\u003c/em\u003e \u003cb\u003e54\u003c/b\u003e (4), 321\u0026ndash;325 (2001).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGrasso, M. et al. The mitochondrial DNA mutation T12297C affects a highly conserved nucleotide of tRNA\u0026thinsp;\u0026lt;\u0026thinsp;SUP\u0026thinsp;\u0026gt;\u0026thinsp;Leu(CUN)\u0026thinsp;and is associated with dilated cardiomyopathy. \u003cem\u003eEur. J. Hum. Genet.\u003c/em\u003e \u003cb\u003e9\u003c/b\u003e (4), 311\u0026ndash;315 (2001).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRobbins, R. C. et al. Cardiac transplantation for hypertrophic cardiomyopathy associated with Sengers syndrome. \u003cem\u003eAnn. Thorac. Surg.\u003c/em\u003e \u003cb\u003e60\u003c/b\u003e (5), 1425\u0026ndash;1427 (1995).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLucas, D. T., Aryal, P., Szweda, L. I., Koch, W. J. \u0026amp; Leinwand, L. A. Alterations in mitochondrial function in a mouse model of hypertrophic cardiomyopathy. \u003cem\u003eAm. J. Physiol. Heart Circ. Physiol.\u003c/em\u003e \u003cb\u003e284\u003c/b\u003e (2), H575\u0026ndash;H83 (2003).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTardiff, J. C. et al. Cardiac troponin T mutations result in allele-specific phenotypes in a mouse model for hypertrophic cardiomyopathy. \u003cem\u003eJ. Clin. Invest.\u003c/em\u003e \u003cb\u003e104\u003c/b\u003e (4), 469\u0026ndash;481 (1999).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUnno, K. et al. Relation of functional and morphological changes in mitochondria to myocardial contractile and relaxation reserves in asymptomatic to mildly symptomatic patients with hypertrophic cardiomyopathy. \u003cem\u003eEur. Heart J.\u003c/em\u003e \u003cb\u003e30\u003c/b\u003e (15), 1853\u0026ndash;1862 (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRanjbarvaziri, S. et al. Altered Cardiac Energetics and Mitochondrial Dysfunction in Hypertrophic Cardiomyopathy. \u003cem\u003eCirculation\u003c/em\u003e \u003cb\u003e144\u003c/b\u003e (21), 1714\u0026ndash;1731 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, H. C. et al. Scientific discovery in the age of artificial intelligence. \u003cem\u003eNature\u003c/em\u003e \u003cb\u003e620\u003c/b\u003e (7972), 47\u0026ndash;60 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang, T., Gradus, J. L. \u0026amp; Rosellini, A. J. Supervised Machine Learning: A Brief Primer. \u003cem\u003eBehav. Ther.\u003c/em\u003e \u003cb\u003e51\u003c/b\u003e (5), 675\u0026ndash;687 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, J., Li, J., Wang, H. \u0026amp; Yan, J. Application of deep learning in genomics. \u003cem\u003eSci. China Life Sci.\u003c/em\u003e \u003cb\u003e63\u003c/b\u003e (12), 1860\u0026ndash;1878 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKiselev, V. Y., Andrews, T. S. \u0026amp; Hemberg, M. Challenges in unsupervised clustering of single-cell RNA-seq data. \u003cem\u003eNat. Rev. Genet.\u003c/em\u003e \u003cb\u003e20\u003c/b\u003e (5), 273\u0026ndash;282 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTenekedjiev, K. I. \u0026amp; Nikolova, N. D. Applications of Monte Carlo simulation in modelling of biochemical processes [Book Chapter]2011.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFaouzi, J. \u0026amp; Colliot, O. Classic machine learning methods arXiv.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, Y. J. et al. Incorporating machine learning and PPI networks to identify mitochondrial fission-related immune markers in abdominal aortic aneurysms. \u003cem\u003eHeliyon\u003c/em\u003e \u003cb\u003e10\u003c/b\u003e (7), e27989 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePickering, J. W. Machine learning for decision-making in cardiology: a narrative review to aid navigating the new landscape. \u003cem\u003eRev. Esp. Cardiol.\u003c/em\u003e \u003cb\u003e76\u003c/b\u003e (8), 645\u0026ndash;654 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFeeny, A. K. et al. Artificial Intelligence and Machine Learning in Arrhythmias and Cardiac Electrophysiology. \u003cem\u003eCirculation-Arrhythmia Electrophysiol.\u003c/em\u003e ;\u003cb\u003e13\u003c/b\u003e(8). (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJohnson, K. W. et al. Artificial Intelligence in Cardiology. \u003cem\u003eJ. Am. Coll. Cardiol.\u003c/em\u003e \u003cb\u003e71\u003c/b\u003e (23), 2668\u0026ndash;2679 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCosta-Silva, J., Domingues, D. \u0026amp; Lopes, F. M. RNA-Seq differential expression analysis: An extended review and a software tool. \u003cem\u003ePlos One\u003c/em\u003e ;\u003cb\u003e12\u003c/b\u003e(12). (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSu, X. X. et al. Bioinformatics and machine learning approaches reveal key genes and underlying molecular mechanisms of atherosclerosis: A review. \u003cem\u003eMedicine\u003c/em\u003e ;\u003cb\u003e103\u003c/b\u003e(31). (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao, W. et al. Weighted Gene Coexpression Network Analysis: State of the Art. \u003cem\u003eJ. Biopharm. Stat.\u003c/em\u003e \u003cb\u003e20\u003c/b\u003e (2), 281\u0026ndash;300 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu, M. M. et al. Identification and validation of immune and oxidative stress-related diagnostic markers for diabetic nephropathy by WGCNA and machine learning. \u003cem\u003eFront. Immunol.\u003c/em\u003e ;\u003cb\u003e14\u003c/b\u003e. (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChicco, D. \u0026amp; Jurman, G. A brief survey of tools for genomic regions enrichment analysis. \u003cem\u003eFront. Bioinf.\u003c/em\u003e ;\u003cb\u003e2\u003c/b\u003e. (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKanehisa, M., Furumichi, M., Sato, Y., Matsuura, Y. \u0026amp; Ishiguro-Watanabe, M. KEGG: biological systems database as a model of the real world. \u003cem\u003eNucleic Acids Res.\u003c/em\u003e \u003cb\u003e53\u003c/b\u003e (D1), D672\u0026ndash;d7 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu, G., Wang, L. G., Han, Y. \u0026amp; He, Q. Y. clusterProfiler: an R package for comparing biological themes among gene clusters. \u003cem\u003eOmics\u003c/em\u003e \u003cb\u003e16\u003c/b\u003e (5), 284\u0026ndash;287 (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePetitprez, F. et al. Transcriptomic analysis of the tumor microenvironment to guide prognosis and immunotherapies. \u003cem\u003eCancer Immunol. Immunotherapy\u003c/em\u003e. \u003cb\u003e67\u003c/b\u003e (6), 981\u0026ndash;988 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAran, D., Hu, Z. C. \u0026amp; Butte, A. J. xCell: digitally portraying the tissue cellular heterogeneity landscape. \u003cem\u003eGenome Biol.\u003c/em\u003e ;\u003cb\u003e18\u003c/b\u003e. (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, S. N. et al. Intratumoral microbial heterogeneity affected tumor immune microenvironment and determined clinical outcome of HBV-related HCC. \u003cem\u003eHepatology\u003c/em\u003e \u003cb\u003e78\u003c/b\u003e (4), 1079\u0026ndash;1091 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu, H. Z. et al. Identification and analysis of key hypoxia- and immune-related genes in hypertrophic cardiomyopathy. \u003cem\u003eBiol. Res.\u003c/em\u003e ;\u003cb\u003e56\u003c/b\u003e(1). (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen, B., Khodadoust, M. S., Liu, C. L., Newman, A. M. \u0026amp; Alizadeh, A. A. Profiling Tumor Infiltrating Immune Cells with CIBERSORT. In: VonStechow L, editor. Cancer Systems Biology: Methods and Protocols. Methods in Molecular Biology. 17112018. pp. 243\u0026thinsp;\u0026ndash;\u0026thinsp;59.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBinson, V. A. et al. A Review of Machine Learning Algorithms for Biomedical Applications. \u003cem\u003eAnn. Biomed. Eng.\u003c/em\u003e \u003cb\u003e52\u003c/b\u003e (4), 1051\u0026ndash;1066 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRosati, D. et al. Differential gene expression analysis pipelines and bioinformatic tools for the identification of specific biomarkers: A review. \u003cem\u003eComput. Struct. Biotechnol. J.\u003c/em\u003e \u003cb\u003e23\u003c/b\u003e, 1154\u0026ndash;1168 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArbet, J., McGue, M., Chatterjee, S. \u0026amp; Basu, S. Resampling-based tests for Lasso in genome-wide association studies. \u003cem\u003eBMC Genet.\u003c/em\u003e ;\u003cb\u003e18\u003c/b\u003e. (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim, Y., Hao, J., Mallavarapu, T., Park, J. \u0026amp; Kang, M. Hi-LASSO: High-Dimensional LASSO. \u003cem\u003eIEEE Access.\u003c/em\u003e \u003cb\u003e7\u003c/b\u003e, 44562\u0026ndash;44573 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSanz, H., Valim, C., Vegas, E., Oller, J. M. \u0026amp; Reverter, F. SVM-RFE: selection and visualization of the most relevant features through non-linear kernels. \u003cem\u003eBMC Bioinform.\u003c/em\u003e ;\u003cb\u003e19\u003c/b\u003e. (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKavitha, K. R., Rajendran, G. S. \u0026amp; Varsha, J. (eds) A Correlation Based SVM-Recursive Multiple Feature Elimination Classifier for Breast Cancer Disease using Microarray. International Conference on Advances in Computing, Communications and Informatics (ICACCI); Sep 21\u0026ndash;24; Jaipur, INDIA2016. (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eD\u0026iacute;az-Uriarte, R. \u0026amp; Alvarez de Andr\u0026eacute;s, S. Gene selection and classification of microarray data using random forest. \u003cem\u003eBMC Bioinform.\u003c/em\u003e \u003cb\u003e7\u003c/b\u003e, 3 (2006).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, J., Malley, J. D., Andrew, A. S., Karagas, M. R. \u0026amp; Moore, J. H. Detecting gene-gene interactions using a permutation-based random forest method. \u003cem\u003eBioData Min.\u003c/em\u003e ;\u003cb\u003e9\u003c/b\u003e. (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMotsinger-Reif, A. A., Deodhar, S., Winham, S. J. \u0026amp; Hardison, N. E. Grammatical evolution decision trees for detecting gene-gene interactions. \u003cem\u003eBioData Min.\u003c/em\u003e ;\u003cb\u003e3\u003c/b\u003e. (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYeh, J. Y. \u0026amp; Wu, T. H. Cascade of genetic algorithm and decision tree for cancer classification on gene expression data. \u003cem\u003eExpert Syst.\u003c/em\u003e \u003cb\u003e27\u003c/b\u003e (3), 201\u0026ndash;218 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, W., Yin, Y. B., Quan, X. W. \u0026amp; Zhang, H. Gene Expression Value Prediction Based on XGBoost Algorithm. \u003cem\u003eFront. Genet.\u003c/em\u003e ;\u003cb\u003e10\u003c/b\u003e. (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang, L. Z., Chen, Y. W. \u0026amp; Huang, W. Hub Genes Identification, Small Molecule Compounds Prediction for Atrial Fibrillation and Diagnostic Model Construction Based on XGBoost Algorithm. \u003cem\u003eFront. Cardiovasc. Med.\u003c/em\u003e ;\u003cb\u003e9\u003c/b\u003e. (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuo, Y. J. et al. Gene-Based Testing of Interactions Using XGBoost in Genome-Wide Association Studies. \u003cem\u003eFront. Cell. Dev. Biol.\u003c/em\u003e ;\u003cb\u003e9\u003c/b\u003e. (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, L. Z. et al. Transformer fault diagnosis method based on SMOTE and NGO-GBDT. \u003cem\u003eSci. Rep.\u003c/em\u003e ;\u003cb\u003e14\u003c/b\u003e(1). (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, Y. et al. A feature selection and multi-model fusion-based approach of predicting air quality. \u003cem\u003eISA Trans.\u003c/em\u003e \u003cb\u003e100\u003c/b\u003e, 210\u0026ndash;220 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePeng, C., Zhang, Y. X., Lang, X. Y. \u0026amp; Zhang, Y. Role of mitochondrial metabolic disorder and immune infiltration in diabetic cardiomyopathy: new insights from bioinformatics analysis. \u003cem\u003eJ. Translational Med.\u003c/em\u003e ;\u003cb\u003e21\u003c/b\u003e(1). (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao, Y. C., Chen, X. Q., Chen, J. J. \u0026amp; Qi, X. Decoding Connectivity Map-based drug repurposing for oncotherapy. \u003cem\u003eBrief. Bioinform.\u003c/em\u003e ;\u003cb\u003e24\u003c/b\u003e(3). (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMusa, A. et al. A review of connectivity map and computational approaches in pharmacogenomics. \u003cem\u003eBrief. Bioinform.\u003c/em\u003e \u003cb\u003e19\u003c/b\u003e (3), 506\u0026ndash;523 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLamb, J. The Connectivity Map: a new tool for biomedical research. \u003cem\u003eNat. Rev. Cancer\u003c/em\u003e. \u003cb\u003e7\u003c/b\u003e (1), 54\u0026ndash;60 (2007).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJi, L. Y. et al. Identification of bioactive compounds and potential mechanisms of scutellariae radix-coptidis rhizoma in the treatment of atherosclerosis by integrating network pharmacology and experimental validation. \u003cem\u003eBiomed. Pharmacother.\u003c/em\u003e ;\u003cb\u003e165\u003c/b\u003e. (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOmmen, S. R. \u0026amp; Nishimura, R. A. Hypertrophic cardiomyopathy. \u003cem\u003eCurr. Probl. Cardiol.\u003c/em\u003e \u003cb\u003e29\u003c/b\u003e (5), 239\u0026ndash;291 (2004).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUllmann, T., Luckhardt, S., Wolf, M., Parnham, M. J. \u0026amp; Resch, E. High-Throughput Screening for CEBPD-Modulating Compounds in THP-1-Derived Reporter Macrophages Identifies Anti-Inflammatory HDAC and BET Inhibitors. \u003cem\u003eInt. J. Mol. Sci.\u003c/em\u003e ;\u003cb\u003e22\u003c/b\u003e(6). (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWedel, A. \u0026amp; Ziegler-Heitbrock, H. W. L. The C/EBP Family of Transcription Factors. \u003cem\u003eImmunobiology\u003c/em\u003e \u003cb\u003e193\u003c/b\u003e (2\u0026ndash;4), 171\u0026ndash;185 (1995).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAkasaka, T. et al. Five members of the CEBP transcription factor family are targeted by recurrent IGH translocations in B-cell precursor acute lymphoblastic leukemia (BCP-ALL). \u003cem\u003eBlood\u003c/em\u003e \u003cb\u003e109\u003c/b\u003e (8), 3451\u0026ndash;3461 (2007).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChan, T. C., Shiue, Y. L. \u0026amp; Li, C. F. The biological impacts of CEBPD on urothelial carcinoma development and progression. \u003cem\u003eFront. Oncol.\u003c/em\u003e ;\u003cb\u003e13\u003c/b\u003e. (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSpek, C. A., Aberson, H. L., Butler, J. M., de Vos, A. F. \u0026amp; Duitman, J. CEBPD Potentiates the Macrophage Inflammatory Response but CEBPD Knock-Out Macrophages Fail to Identify CEBPD-Dependent Pro-Inflammatory Transcriptional Programs. \u003cem\u003eCells\u003c/em\u003e ;\u003cb\u003e10\u003c/b\u003e(9). (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKo, C. Y., Chang, W. C. \u0026amp; Wang, J. M. Biological roles of CCAAT/Enhancer-binding protein delta during inflammation. \u003cem\u003eJ. Biomed. Sci.\u003c/em\u003e ;\u003cb\u003e22\u003c/b\u003e. (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUllmann, T., Luckhardt, S., Wolf, M., Parnham, M. J. \u0026amp; Resch, E. High-Throughput Screening for CEBPD-Modulating Compounds in THP-1-Derived Reporter Macrophages Identifies Anti-Inflammatory HDAC and BET Inhibitors. \u003cem\u003eInt. J. Mol. Sci.\u003c/em\u003e ;\u003cb\u003e22\u003c/b\u003e(6). (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, J. M., Ko, C. Y., Chen, L. C., Wang, W. L. \u0026amp; Chang, W. C. Functional role of NF-IL6beta and its sumoylation and acetylation modifications in promoter activation of cyclooxygenase 2 gene. \u003cem\u003eNucleic Acids Res.\u003c/em\u003e \u003cb\u003e34\u003c/b\u003e (1), 217\u0026ndash;231 (2006).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCho, M. K., Cho, Y. H., Lee, G. H. \u0026amp; Kim, S. G. Induction of cyclooxygenase-2 by bovine type I collagen in macrophages via C/EBP and CREB activation by multiple cell signaling pathways. \u003cem\u003eBiochem. Pharmacol.\u003c/em\u003e \u003cb\u003e67\u003c/b\u003e (12), 2239\u0026ndash;2250 (2004).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThomas, B. et al. Critical role of C/EBPdelta and C/EBPbeta factors in the stimulation of the cyclooxygenase-2 gene transcription by interleukin-1beta in articular chondrocytes. \u003cem\u003eEur. J. Biochem.\u003c/em\u003e \u003cb\u003e267\u003c/b\u003e (23), 6798\u0026ndash;6809 (2000).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBalamurugan, K. et al. FBXW7α attenuates inflammatory signalling by downregulating C/EBPδ and its target gene Tlr4. \u003cem\u003eNat. Commun.\u003c/em\u003e ;\u003cb\u003e4\u003c/b\u003e. (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, Y. H. et al. CEBPD amplification and overexpression in urothelial carcinoma: a driver of tumor metastasis indicating adverse prognosis. \u003cem\u003eOncotarget\u003c/em\u003e \u003cb\u003e6\u003c/b\u003e (31), 31069\u0026ndash;31084 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTsai, H. H. et al. Metformin promotes apoptosis in hepatocellular carcinoma through the CEBPD-induced autophagy pathway. \u003cem\u003eOncotarget\u003c/em\u003e \u003cb\u003e8\u003c/b\u003e (8), 13832\u0026ndash;13845 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, Q. W. et al. A hierarchical and collaborative BRD4/CEBPD partnership governs vascular smooth muscle cell inflammation. \u003cem\u003eMol. Therapy-Methods Clin. Dev.\u003c/em\u003e \u003cb\u003e21\u003c/b\u003e, 54\u0026ndash;66 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChi, J. Y. et al. Fibroblast CEBPD/SDF4 axis in response to chemotherapy-induced angiogenesis through CXCR4. \u003cem\u003eCell. Death Discovery\u003c/em\u003e ;\u003cb\u003e7\u003c/b\u003e(1). (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChan, T. C. et al. Biological significance of MYC and CEBPD coamplification in urothelial carcinoma: Multilayered genomic, transcriptional and posttranscriptional positive feedback loops enhance oncogenic glycolysis. \u003cem\u003eClin. Transl Med.\u003c/em\u003e \u003cb\u003e11\u003c/b\u003e (12), e674 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, W. J. et al. MCL1 participates in leptin-promoted mitochondrial fusion and contributes to drug resistance in gallbladder cancer. \u003cem\u003eJCI Insight\u003c/em\u003e ;\u003cb\u003e6\u003c/b\u003e(15). (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBanerjee, S. et al. Loss of C/EBPδ enhances IR-induced cell death by promoting oxidative stress and mitochondrial dysfunction. \u003cem\u003eFree Radic Biol. Med.\u003c/em\u003e \u003cb\u003e99\u003c/b\u003e, 296\u0026ndash;307 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuha, M., Srinivasan, S., Koenigstein, A., Zaidi, M. \u0026amp; Avadhani, N. G. Enhanced osteoclastogenesis by mitochondrial retrograde signaling through transcriptional activation of the cathepsin K gene. \u003cem\u003eAnn. N Y Acad. Sci.\u003c/em\u003e \u003cb\u003e1364\u003c/b\u003e (1), 52\u0026ndash;61 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eL\u0026ouml;ffler, M. C. et al. Protein kinase D1 deletion in adipocytes enhances energy dissipation and protects against adiposity. \u003cem\u003eEmbo j.\u003c/em\u003e ;\u003cb\u003e37\u003c/b\u003e(22). (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOmmen, S. R. et al. 2020 AHA/ACC Guideline for the Diagnosis and Treatment of Patients With Hypertrophic Cardiomyopathy: Executive Summary A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. \u003cem\u003eJ. Am. Coll. Cardiol.\u003c/em\u003e \u003cb\u003e76\u003c/b\u003e (25), 3022\u0026ndash;3055 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMaron, B. J., Maron, M. S., Maron, B. A. \u0026amp; Loscalzo, J. Moving Beyond the Sarcomere to Explain Heterogeneity in Hypertrophic Cardiomyopathy JACC Review Topic of the Week. \u003cem\u003eJ. Am. Coll. Cardiol.\u003c/em\u003e \u003cb\u003e73\u003c/b\u003e (15), 1978\u0026ndash;1986 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCarrillo-Salinas, F. J., Ngwenyama, N., Anastasiou, M., Kaur, K. \u0026amp; Alcaide, P. Heart Inflammation Immune Cell Roles and Roads to the Heart. \u003cem\u003eAm. J. Pathol.\u003c/em\u003e \u003cb\u003e189\u003c/b\u003e (8), 1482\u0026ndash;1494 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOzyilmaz, S. et al. The importance of the neutrophil-to-lymphocyte ratio in patients with hypertrophic cardiomyopathy. \u003cem\u003eRev. Port. Cardiol.\u003c/em\u003e \u003cb\u003e36\u003c/b\u003e (4), 239\u0026ndash;246 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZheng, X. F., Yang, Y., Fu, C. H. \u0026amp; Huang, R. A. Identification and verification of promising diagnostic biomarkers in patients with hypertrophic cardiomyopathy associate with immune cell infiltration characteristics. \u003cem\u003eLife Sci.\u003c/em\u003e ;\u003cb\u003e285\u003c/b\u003e. (2021).\u003c/span\u003e\u003c/li\u003e \u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Hypertrophic cardiomyopathy, immune microenvironment, machine learning, mitochondrial dysfunction, molecular docking, bioinformatics","lastPublishedDoi":"10.21203/rs.3.rs-5122992/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5122992/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cb\u003ePurpose\u003c/b\u003e\u003c/p\u003e \u003cp\u003eThis study looked at possible targets for hypertrophic cardiomyopathy (HCM), a condition marked by thickening of the ventricular wall, primarily in the left ventricle.\u003c/p\u003e\u003cp\u003e\u003cb\u003eMethods\u003c/b\u003e\u003c/p\u003e \u003cp\u003eWe employed differential gene analysis and weighted gene co-expression network analysis (WGCNA) on samples. We then carried out an enrichment analysis. We also investigated the process of immunological infiltration. We employed six machine learning techniques and two protein-protein interaction (PPI) network gene selection approaches to search for the most characteristic gene (MCG). In the validation ladder, we verified the expression of MCG. Furthermore, we examined the MCG expression levels in HCM animal and cell models. Finally, we performed molecular docking and predicted potential medications for HCM treatment.\u003c/p\u003e\u003cp\u003e\u003cb\u003eResults\u003c/b\u003e\u003c/p\u003e \u003cp\u003e7975 differentially expressed genes (DEGs) were found in our study. We also identified 236 genes in the blue module using WGCNA. Screening at the transcriptome and protein levels was used to mine MCG. The final result screened CCAAT/Enhancer Binding Protein Delta (CEBPD) as MCG. We confirmed that MCG expression matched the outcomes of the experimental ladder. The level of CEBPD mRNA and protein was lowered in HCM animal and cellular models. Given that Abt-751 had the highest binding affinity to CEBPD, it might be a projected targeted medication.\u003c/p\u003e\u003cp\u003e\u003cb\u003eConclusion\u003c/b\u003e\u003c/p\u003e \u003cp\u003eWe found a new target gene for HCM called CEBPD, which is probably going to function by mitochondrial dysfunction. An innovative aim for the management or avoidance of HCM is offered by this analysis. Abt-751 may be a predicted targeted drug for HCM that had the greatest binding affinity with CEBPD.\u003c/p\u003e","manuscriptTitle":"Protein interactions, network pharmacology, and machine learning work together to predict genes linked to mitochondrial dysfunction in hypertrophic cardiomyopathy","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-03-24 09:46:29","doi":"10.21203/rs.3.rs-5122992/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Accepted","date":"2025-04-04T15:46:09+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-04-04T07:29:22+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-03-31T23:19:35+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"143845771530164996056189106574766546353","date":"2025-03-24T17:41:25+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"261318485565956223979697730755311591664","date":"2025-03-21T15:24:56+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-03-21T12:15:24+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-03-21T09:47:21+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-03-14T12:56:17+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"8795cd97-12a2-4862-b46e-8d3363cce352","owner":[],"postedDate":"March 24th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":46035404,"name":"Biological sciences/Biological techniques"},{"id":46035405,"name":"Biological sciences/Biotechnology"},{"id":46035406,"name":"Biological sciences/Computational biology and bioinformatics"},{"id":46035407,"name":"Biological sciences/Immunology"}],"tags":[],"updatedAt":"2025-05-05T15:58:50+00:00","versionOfRecord":{"articleIdentity":"rs-5122992","link":"https://doi.org/10.1038/s41598-025-97534-x","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2025-04-29 15:57:05","publishedOnDateReadable":"April 29th, 2025"},"versionCreatedAt":"2025-03-24 09:46:29","video":"","vorDoi":"10.1038/s41598-025-97534-x","vorDoiUrl":"https://doi.org/10.1038/s41598-025-97534-x","workflowStages":[]},"version":"v1","identity":"rs-5122992","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5122992","identity":"rs-5122992","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}