Novel drug resistance- and macrophage polarization-related molecular subtyping and prognostic signature for pancreatic adenocarcinoma

Novel drug resistance- and macrophage polarization-related molecular subtyping and prognostic signature for pancreatic adenocarcinoma | 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 Research Article Novel drug resistance- and macrophage polarization-related molecular subtyping and prognostic signature for pancreatic adenocarcinoma Zixin Liu, Ruijiao Kong, Zichen Liu, Yin Jia, Gang Jin This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7991235/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Backgroud: Pancreatic adenocarcinoma (PAAD) is characterized by an aggressive behavior and poor prognosis, requiring innovative therapeutic strategies. Methods: The PAAD datasets were acquired from two publicly available genomic repositories: The Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus (GEO) database. Drug resistance- and macrophage polarization-related genes (DMRGs) were collected on GeneCards or DRESIS databases. To identify distinct disease subtypes, we identifiedprognostic genes with univarite COX regression analysis, followed by consensus clustering. Then an intersection analysis between differentially expressed genes (DEGs) and a set of DMRGs was performed, and the overlapping genes yielded drug resistance- and macrophage polarization-related differentially expressed genes (DMRDEGs). Based on DMRDEGs identified, a prognostic risk model was constructed. Results: PAAD patients were categorized into two molecularly distinct subgroups, subtype A (1) and subtype B (2), based on DMRGs. Through immunological profiling, we found five distinct immune cell populations with statistically significant variations, notably comprising regulatory T lymphocytes and activated NK cells. Immunological profiling demonstrated that subtype B displayed increased sensitivity to immunotherapy (p 0.7). A protein-protein interaction (PPI) network was established focusing on these genes, revealing their function as key regulatory hubs. Conclusion: Our analysis categorized PAAD into two distinct subgroups based on DMRGs and a prognostic risk model developed from these genes exhibits considerable promise for forecasting patient survival outcomes. pancreatic adenocarcinoma drug resistance macrophage polarization molecular characterization immune microenvironment Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1 Introduction Pancreatic adenocarcinoma (PAAD) is recognized as a highly aggressive form of solid malignancy, characterized by a notably poor clinical outcome. Current treatment modalities for PAAD are limited and primarily include surgical resection, chemotherapy, radiotherapy, and immunotherapy [ 1 ]. However, the efficacy of these approaches is often compromised due to the emergence of drug resistance. Therefore, identifying novel biomarkers is critical for stratifying PAAD patients and assessing the prognosis of distinct patient subgroups [ 2 ]. The PAAD tumor microenvironment (TME) exhibits a dense extracellular matrix enriched with activated fibroblasts and immune cells [ 3 ]. A dense stroma impairs vascularization and vessel functionality, thereby hindering the effective delivery of therapeutic agents. Tumor-associated macrophages (TAMs), as pivotal cellular constituents within the tumor microenvironment (TME), are key players in the growth, recurrence, and spread of PAAD tumors and in modulating immune responses and reactions to chemotherapy and immunotherapy [ 4 ]. Elevated infiltration of TAMs are in connection with poor therapeutic outcomes in PAAD, and reducing the number of TAMs has been shown to suppress pancreatic tumor development [ 5 ]. Nevertheless, the intrinsic factors governing macrophage development and function in PAAD remain largely unknown, which significantly limits the therapeutic potential of TAMs as targets and tools for PAAD treatment. Prior studies have emphasized the fundamental importance of DMRGs in modulating the tumor microenvironment and contributing to therapeutic resistance [ 6 ]. Despite these advances, a significant gap remains in the identification of distinct pancreatic cancer subtypes and development of robust prognostic models that integrate these genetic factors. Therefore, this study discovered novel biomarkers and therapeutic targets to enhance prognostic accuracy and ultimately improve clinical outcomes in the management of PAAD. To this end, we employed a comprehensive bioinformatics approach incorporating techniques such as batch effect removal, consistent clustering analysis, differential expression, and Cox regression modeling. The advantage of this multifaceted approach is its ability to integrate information from multiple datasets, thereby enhancing the reliability and reproducibility of the results. The primary objective of this study was to identify the distinct subtypes of pancreatic cancer and investigate the expression characteristics of DMRGs. Through the development of a comprehensive prognostic scoring system, we aimed to elucidate the pivotal roles of these molecular markers in PAAD. This approach facilitates the discovery of novel diagnostic biomarkers and potential therapeutic interventions, ultimately improving patient prognosis and treatment efficacy. 2 Materials and methods 2.1 Data download The PAAD dataset (TCGA-PAAD), was obtained from TCGA database and utilized for analytical evaluation. Samples lacking clinical information were excluded and then we obtained sequencing data in count format for 176 PAAD samples with clinical information. Meanwhile, the data was accessed through the UCSC Xena database to acquire corresponding clinical data (Table 1 ). Table 1 Baseline Table with PAAD Patients Characteristics characteristics overall Gender, n (%) FEMALE 79 (44.9%) MALE 97 (55.1%) Stage_M, n (%) M0 79 (44.9%) M1&MX 97 (55.1%) Stage_N, n (%) N0 49 (27.8%) N1&NX 127 (72.2%) Stage_T, n (%) T1&T2 30 (17%) T3&T4&TX 146 (83%) PAAD, Pancreatic adenocarcinoma The PAAD datasets GSE28735 and GSE71729 were obtained from the Gene Expression Omnibus (GEO) database using the GEOquery package (version 2.70.0) [ 7 ]. Dataset GSE28735 uses the GPL6244 chip platform, while dataset GSE71729 utilizes GPL20769. More details are available in Table 2 . The dataset GSE28735 contains 45 PAAD and 45 control samples. The dataset GSE71729 includes 145 PAAD and 46 control samples. Table 2 GEO Microarray Chip Information GSE28735 GSE71729 Platform GPL6244 GPL20769 Species Homo sapiens Homo sapiens Tissue Pancreas Pancreas Samples in PAAD group 45 145 Samples in Control group 45 46 Reference PMID: 22363658 PMID: 23918603 PMID: 26343385 GEO, Gene Expression Omnibus;PAAD, Pancreatic adenocarcinoma Through comprehensive mining of the GeneCards [ 8 ] and DRESIS [ 9 ] databases, we compiled a catalog of genes implicated in both drug resistance mechanisms and macrophage polarization processes (hereafter referred to as DMRGs). We first downloaded the drug resistance-related gene set from the DRESIS database, retaining only the drug resistance-related genes (DRGs) with molecule type as 'Protein' and Molecule species as 'Homo sapiens.’ A systematic literature search was conducted on PubMed using "drug resistance" as the primary search term to identify relevant published studies [ 10 , 11 ]. The set of DRGs was consolidated and de-duplicated, yielding 959 DRGs. Similarly, we identified 586 macrophage polarization-related genes (MRGs) [ 12 – 15 ]. After determining the intersection of genes related to the two phenotypes of drug resistance and macrophage polarization, 106 DMRGs were obtained (Supplementary material 4 - Table S1 ). The batch effect correction was implemented on GSE28735 and GSE71729 datasets through the sva package (v3.50.0), resulting in a consolidated GEO dataset containing 190 PAAD specimens. To evaluate the normalization efficacy, expression profile distributions were visualized and compared using boxplot representations prior to and following the standardization procedure. 2.2 Construction of PAAD subtypes To identify potential prognostic biomarkers in PAAD from the TCGA dataset, univariate Cox proportional hazards regression was conducted incorporating clinical parameters. Following this initial screening, we employed consensus clustering methodology to stratify PAAD cases into molecularly distinct subgroups characterized by differential expression patterns of the identified prognostic genes. 2.3 DEGs analysis Based on the grouping of PAAD samples from TCGA-PAAD, gene expression analysis was conducted across various disease subtypes. Differential gene expression analysis identified statistically significant genes (DEGs) using stringent criteria: an absolute log2 fold change greater than 1 (|log2FC| > 1) combined with an adjusted p-value below 0.05. An intersection analysis between DEGs and a set of DMRGs was performed and the overlapping genes yielded the DMRDEGs. The 20 most significant differentially DMRDEGs were visualized through pheatmap package (v1.0.12) in R. Furthermore, their genomic distribution across chromosomes was mapped and presented using the RCircos package (v1.2.2). 2.4 Prognosis analysis of pancreatic cancer Kaplan-Meier (KM) survival analysis was performed to assess survival outcome disparities between different molecular subtypes of PAAD. Additionally, the ggplot2 package was utilized to generate visual representations of clinical feature distributions across patient subgroups, enabling systematic evaluation of potential prognostic determinants. 2.5 Analysis of somatic mutations (SMs) and copy number variations (CNVs) For SM data processing, we employed VarScan for initial quality control and variant calling. Subsequently, we performed comprehensive characterization of CNV patterns, including their genomic distribution and prevalence, through visualization analyses implemented in ggplot2 (v3.4.4). 2.6 Immunotherapy analysis We sourced 25 immune checkpoint genes (ICGs) from the PubMed literature [ 16 , 17 ], and their names are detailed in Table S2 (Supplementary material 5). The expression differences of these ICGs in PAAD samples were analyzed, and grouped comparison charts were generated. The immune infiltration scores for PAAD samples were calculated using transcriptomic data obtained from the TCGA-PAAD cohort through the TIDE online platform [ 18 ]. 2.7 Gene set enrichment analysis (GSEA) GSEA was performed on the comprehensive RNA sequencing dataset obtained from TCGA-PAAD cohort, utilizing version 4.10.0 of the clusterProfiler R package. The analytical parameters were carefully selected to optimize both statistical power and biological significance: a fixed random seed (2020) was implemented to guarantee reproducibility, 1000 permutations were executed to establish robust statistical inference, and gene sets were restricted to those comprising between 10 and 500 genes to maintain biological interpretability. Pathway annotations were obtained from the Molecular Signatures Database (MSigDB) [ 19 ]. Statistically significant pathways were determined using strict criteria: Benjamini-Hochberg corrected p-values less than 0.05 and false discovery rates (q-values) below 0.25, which provided robust control for multiple comparisons. 2.8 Development of a predictive risk stratification model for PAAD Initially, we recognized prognostic factors demonstrating statistical significance (p < 0.05). These significant variables were subjected into a model for further evaluation. Subsequently, to construct the prognostic model, we employed Least Absolute Shrinkage and Selection Operator (LASSO) regression analysis with the family parameter specified as "cox". This analytical approach utilized DMRDEGs identified in the previous analysis. Through 10-fold cross-validation, we determined the optimal set of predictive genes for the risk assessment model. The results of this LASSO regression analysis were visually represented through two distinct graphical outputs: a risk score prognostic plot and a coefficient profile trajectory plot. 2.9 Evaluation of PAAD risk prediction model The predictive model's ability to distinguish temporal patterns was evaluated by conducting time-dependent receiver operating characteristic (ROC) curve analysis, implemented using the survivalROC package (version 1.0.3.1), which enabled the computation of area under the curve (AUC) values for 1-, 2-, and 3-year survival predictions. Survival analysis was conducted using the survival package (v3.5-7) to generate Kaplan-Meier curves. For clinical application, we developed a prognostic nomogram through the rms package (v6.7-1), incorporating significant covariates identified. 2.10 Immune infiltration analysis The differential expression patterns of immune cells across subtypes (Cluster 1 vs. Cluster 2) were visualized using ggplot2 (v3.4.4). For comprehensive representation of correlation analyses, pheatmap (v1.0.12) facilitated the generation of heatmaps depicting both intercellular relationships among immune cells and associations between model genes and immune cell populations. This dual visualization approach enabled systematic examination of immune microenvironment characteristics. 2.11 Validation of differential expression and analysis of ROC Curve We determined risk scores for each PAAD sample across the integrated GEO datasets based on established risk coefficients. A comparative comparison chart was generated to visualize the differential expression patterns of the model genes across the two risk groups. We constructed a comparative chart to illustrate the distinct transcriptional profiles of signature genes between high-risk and low-risk patient cohorts. 2.12 Construction of protein-protein interaction networks and identification of hub genes Employing the STRING database [ 20 ], we established a PPI network comprising model genes and subsequently identified candidate genes from this network for additional investigation. Furthermore, the GeneMANIA web-based tool [ 21 ] was utilized to detect genes exhibiting functional associations with hub genes implicated in drug resistance and macrophage polarization. 2.13 Construction of regulatory networks To investigate transcription factor-mediated (TF-mediated) gene regulation, we systematically examined the regulatory interactions between transcription factors and their target genes using data obtained from the ChIPBase repository [ 22 ]. we obtained miRNAs associated with model genes and predicted the target RNA-Binding Protein (RBP) for model genes Using StarBase v3.0 database [ 23 ]. 2.14 ESTIMATE analysis We calculated ESTIMATE immune and stromal scores with R package IOBR (version 0.99.9). And we generated comparative plots for three distinct immune infiltration analyses. 2.15 Statistical analysis For the comparison of continuous variables across two groups, parametric data were assessed via independent two-sample t-tests, whereas nonparametric distributions were analyzed using Mann-Whitney U tests. Bivariate associations were quantified through Spearman's rank correlation coefficients. The complete statistical workflow was implemented using R statistical computing environment (version 4.3.0). 3 Results 3.1 Two subtypes were identified in PAAD based on DMRGs The methodological pipeline for the systematic investigation of DMRGs is depicted in Fig. 1 A. Initially, we eliminated batch effects between the GSE28735 and GSE71729 PAAD datasets with “sva” package, resulting in an integrated GEO dataset. The comparative analysis of expression profiles, as visualized through distribution box plots, revealed distinct patterns between pre- and post-batch correction datasets (Supplementary material 1 - Fig. S1 A and S1B). To further evaluate the intrinsic structure of the integrated dataset, principal component analysis (PCA) was employed to visualize the distribution patterns in reduced dimensions (Supplementary material 1 - Fig. S1 C and S1D). The graphical representations clearly demonstrate the effectiveness of the normalization procedure in reducing technical variability while preserving biological signals. Among the 106 DMRGs, 36 were identified as significantly associated with prognosis based on univariate Cox regression analysis, and their hazard ratios were graphically represented in a forest plot (Fig. 1 B). According to the expression profiles of 36 DMRGs, a consensus clustering analysis was performed to identify distinct molecular subtypes. The analytical results delineated two distinct molecular subgroups of PAAD: Subtype A (Cluster 1), which encompassed 162 tumor specimens, and Subtype B (Cluster 2), containing 14 cases (Figs. 1 C, 1 D, and Supplementary material 2 - Fig. S2 ). This classification was consistently supported by multiple analytical approaches. The three-dimensional t-SNE clustering plot demonstrated distinct clustering patterns that clearly differentiated between the two subtypes (Fig. 1 E). To investigate potential variations in gene expression profiles and their associations with biological properties, disease pathogenesis, and clinical phenotypes, we performed comparative transcriptomic analysis. Applying stringent statistical criteria (|log2 fold change| > 1 with adjusted p-value < 0.05), our differential expression analysis of TCGA-PAAD samples identified 11,176 significantly dysregulated genes. Among these, 4,521 transcripts exhibited increased expression levels while 6,655 showed decreased expression patterns, as illustrated in Fig. 1 F. Through comprehensive analysis, we identified 34 DMRDEGs from the intersection of 11,142 DEGs and 36 DMRGs, as illustrated in Fig. 1 G. Complete annotation data for these candidate genes are provided in Table S3 (Supplementary material 6). Statistical analysis revealed significant differential expression patterns among the top 20 DMRDEGs when ranked according to the magnitude of their log2 fold-change values in the TCGA-PAAD dataset (Fig. 1 H). To conclude our analysis, we employed the RCircos package in R to conduct chromosomal localization studies, which enabled precise mapping of gene loci across the genome. This computational approach yielded a detailed visualization of genomic distribution patterns, as illustrated in Fig. 1 I. And the results indicated that the 34 DMRDEGs mainly located on chromosome 1, specifically, PIK3CD , SLC2A1 , S100A8 , MUC1 , and PTGS2 . 3.2 Characteristics of the two PAAD subtypes Our GSEA analysis demonstrated substantial enrichment across multiple biological processes and signaling cascades (Table 3 and Fig. 2 A). Notably, we observed prominent activation of the Fc epsilon receptor (Fceri) signaling pathway (Fig. 2 B), along with FCGR3A-dependent interleukin-10 ( IL-10 ) production (Fig. 2 C). The study further identified Fceri-triggered calcium ion flux (Fig. 2 D) and the participation of the LAT2/NTAL mechanism in calcium signaling (Fig. 2 E). Table 3 Results of GSEA for PAAD ID Set Size Enrichment Score NES pvalue p.adjust qvalue REACTOME_ROLE_OF_LAT2_NTAL_LAB_ON_CALCIUM_MOBILIZATION 71 0.879869 3.280909 1.00E-10 5.24E-09 3.75E-09 REACTOME_FCERI_MEDIATED_CA_2_MOBILIZATION 86 0.836874 3.244674 1.00E-10 5.24E-09 3.75E-09 REACTOME_FCGR3A_MEDIATED_IL10_SYNTHESIS 94 0.784836 3.106986 1.00E-10 5.24E-09 3.75E-09 REACTOME_FC_EPSILON_RECEPTOR_FCERI_SIGNALING 186 0.657411 2.884516 1.00E-10 5.24E-09 3.75E-09 REACTOME_ASSEMBLY_OF_COLLAGEN_FIBRILS_AND_OTHER_MULTIMERIC_STRUCTURES 61 0.725634 2.672979 1.00E-10 5.24E-09 3.75E-09 WP_OVERVIEW_OF_PROINFLAMMATORY_AND_PROFIBROTIC_MEDIATORS 115 0.654766 2.651993 1.00E-10 5.24E-09 3.75E-09 REACTOME_INTERLEUKIN_10_SIGNALING 45 0.757527 2.618502 1.00E-10 5.24E-09 3.75E-09 REACTOME_INTERLEUKIN_4_AND_INTERLEUKIN_13_SIGNALING 111 0.560062 2.264255 8.85E-10 3.91E-08 2.79E-08 REACTOME_MET_PROMOTES_CELL_MOTILITY 41 0.657049 2.234189 1.10E-06 2.34E-05 1.67E-05 REACTOME_NEUTROPHIL_DEGRANULATION 475 0.463173 2.230004 1.00E-10 5.24E-09 3.75E-09 REACTOME_SIGNALING_BY_INTERLEUKINS 470 0.463849 2.229662 1.00E-10 5.24E-09 3.75E-09 REACTOME_ANTI_INFLAMMATORY_RESPONSE_FAVOURING_LEISHMANIA_PARASITE_INFECTION 211 0.497813 2.211216 1.00E-10 5.24E-09 3.75E-09 REACTOME_TNF_RECEPTOR_SUPERFAMILY_TNFSF_MEMBERS_MEDIATING_NON_CANONICAL_NF_KB_PATHWAY 17 0.79427 2.179953 5.80E-06 0.000104 7.40E-05 REACTOME_DECTIN_2_FAMILY 26 0.703239 2.15809 1.05E-05 0.000171 0.000122 REACTOME_CROSSLINKING_OF_COLLAGEN_FIBRILS 18 0.77486 2.151253 2.82E-05 0.000406 0.00029 PID_AMB2_NEUTROPHILS_PATHWAY 41 0.623443 2.119918 9.98E-06 0.000167 0.000119 PID_PI3KCI_PATHWAY 48 0.59446 2.089981 1.73E-05 0.000271 0.000194 BIOCARTA_IL1R_PATHWAY 31 0.657375 2.087755 3.19E-05 0.00045 0.000322 REACTOME_ANCHORING_FIBRIL_FORMATION 15 0.776326 2.076566 9.39E-05 0.001098 0.000785 REACTOME_COSTIMULATION_BY_THE_CD28_FAMILY 72 0.551565 2.065634 1.16E-06 2.42E-05 1.73E-05 GSEA, Gene Set Enrichment Analysis KM survival analysis revealed that subtypes A (cluster 1) has a highly poorer OS than subtypes B (cluster 2) (p < 0.01, Fig. 2 F). Furthermore, we analyzed the proportions of patients in subtype 1 and subtype 2 across Age, Gender, Stage_M, Stage_N, and Stage_T categories. The figure indicates that more PAAD patients in subtypes A (cluster 1) are in the worse T3, T4 and TX stages than B (cluster 2) (Fig. 2 G). 3.3 Immune environments differ between the two PAAD subtypes Employing the CIBERSORT computational method, we estimated the proportions of 22 distinct immune cell populations within the TCGA-PAAD cohort. Five distinct immune cell populations—regulatory T cells (Tregs), activated natural killer (NK) cells, monocytes, M0 macrophages, and M1 macrophages—exhibited statistically significant variations across the identified subtypes (p < 0.05; Fig. 3 A). Subsequently, a heatmap was employed to illustrate the associations among the infiltration levels of these five immune cell types. The results showed that most immune cells exhibited a strong correlation in these two subtypes, with the strongest significant negative correlation between monocytes and M0 macrophages in subtype A (r = -0.418, p < 0.05; Fig. 3 B), and the robust inverse association between monocytes and M0 macrophages in subtype B (r = -0.55, p < 0.05; Fig. 3 C). ICGs were identified based on previously published studies and publicly accessible databases. Through cross-referencing these genes with the TCGA-PAAD database, we obtained an expression profile matrix containing 25 ICGs and their respective transcriptional quantities. Subsequently, we analyzed the differential expression patterns of ICGs between subtype A (Cluster 1) and subtype B (Cluster 2), as illustrated in Fig. 3 D. Significant differences were observed between subtypes A and B for most ICGs (p < 0.001), including FCRL4, G0S2, CD38, CLEC7A, CCL13, CCR2, CXCL10, CD1E, LILRB2, CD47, CD70, CXCL1, HLA-DOB, CXCL11, LILRB1, CCL18, CCR7, LGALS9 and BIRC3 . Our investigation focused on assessing the responsiveness of PAAD patients to immunotherapeutic interventions, leveraging this computational approach (Fig. 3 E). The results indicated that subtype B may exhibit better responsiveness to immunotherapy than subtype A. 3.4 Analysis of CNVs and SMs in DMRDEGs To investigate SM profiles of the 34 DMRDEGs in PAAD samples from the TCGA-PAAD cohort, we conducted comprehensive mutational profiling using the "maftools" R package. Our analysis revealed seven principal SM categories affecting these DMRDEGs, with missense mutations representing the most prevalent alteration type (Fig. 4 A, left panel). Additionally, the primary mutation type observed among the 34 DMRDEGs within the PAAD group was single nucleotide polymorphisms (SNPs) (Fig. 4 A, middle panel). The predominant single-nucleotide variations (SNVs) identified in PAAD specimens were characterized by cytosine-to-thymine transitions, as demonstrated in the right panel of Fig. 4 A right panel. Furthermore, we analyzed the SM profiles of the 34 DMRDEGs and ranked them based on mutation frequency, followed by visual representation (Fig. 4 B). ANXA1 and PIK3CG exhibited the highest mutation rates and their mutation frequency was 2%. To analyze the CNVs in the 34 DMRDEGs, we downloaded and synthesized CNV data from the PAAD group. Using GISTIC 2.0, CNVs of 34 DMRDEGs in the PAAD group were identified, and the CNV profiles of these genes were visualized (Figs. 4 C and D). 3.5 Development of a predictive risk stratification framework for PAAD A prognostic risk model for PAAD was constructed with 34 DMRDEGs. All significant variables (p-value < 0.05) were visualized using a forest plot (Fig. 5 A). Through comprehensive analysis, 19 DMRDEGs exhibiting significant prognostic relevance (p < 0.05) were identified. To further analysis, a LASSO regression model was developed (Figs. 5 B and 5 C). The final model incorporated five key genes: IL18 , EREG , LDHA , SOCS2 , and SPP1 . The risk score was derived based on the following equation: $$\:RiskScore\:=\:IL18\:\ast\:\:\left(0.112\right)\:+\:EREG\:\ast\:\:\left(0.05\right)\:+\:LDHA\:\ast\:\:\left(0.256\right)\:+\:SOCS2\:\ast\:\:(-0.231)\:+\:SPP1\:\ast\:\:\left(0.028\right)"$$ To investigate the functional associations between the five key genes identified in our model, we conducted a PPI analysis (Fig. 5 D). Building upon these findings, we developed a comprehensive interaction network that integrates these hub genes involved in both drug resistance mechanisms and macrophage polarization processes, along with their functionally related counterparts (Fig. 5 E). The resulting network architecture included the five core model genes and an additional set of 20 proteins sharing similar biological functions. To elucidate the molecular mechanisms governing transcriptional regulation, we constructed a comprehensive mRNA-TF interaction network. This network was subsequently visualized and analyzed using the Cytoscape platform, enabling systematic examination of potential regulatory relationships (Fig. 5 F). This regulatory network incorporated four key model genes along with 36 TF, with comprehensive data available in Supplementary Table S4 (Supplementary material 7). Subsequently, we predicted microRNAs (miRNAs) that potentially interact with these model genes, leading to the construction of an mRNA-miRNA regulatory network. This secondary network encompassed three model genes and 43 miRNAs (Supplementary material 3 - Fig. S3 ), with complete annotation details provided in Supplementary Table S5 (Supplementary material 8). Finally, we constructed an mRNA–RBP regulatory network to visualize RBPs associated with the model genes and the network comprised four model genes and 42 RBPs (Fig. 5 G and Supplementary material 9 - Table S6 ). Furthermore, Statistical evaluation of the association between the predictive gene markers and immune cell infiltration patterns demonstrated that the majority of immune cell types showed significant associations with the five model genes in both subtype A and subtype B. And SOCS2 had the strongest negative correlation with M0 macrophages in subtype A (r = -0.458, p < 0.05; Fig. 5 H), while SOCS2 had the strongest significant positive correlation with activated NK cells in subtype B (r = 0.626, p < 0.05; Fig. 5 I). 3.6 Clinical evaluation of the prognostic risk model Subsequently, we constructed ROC curves utilizing the TCGA-PAAD cohort to evaluate the predictive performance of our prognostic signature (Fig. 6 A). This analytical approach enabled us to quantitatively assess the model's discrimination capacity at various time points during follow-up. The results showed that the model demonstrated moderate to high accuracy at 1 and 3 years (0.7 < AUC < 0.9), with slightly lower accuracy at 2 years (0.5 < AUC < 0.7). KM survival analysis demonstrated a statistically significant divergence in overall survival between high- and low-risk patient cohorts (p < 0.01; Fig. 6 B). The nomogram constructed to represent the association between risk scores and five key clinical variables, indicated that the riskscore exhibited substantially greater predictive value within the prognostic model compared to the other variables (Fig. 6 C). Subsequently, we conducted a univariate Cox proportional hazards regression analysis utilizing the median RiskScore as the stratification threshold, incorporating OS data and relevant clinical parameters extracted from the database. The results from both univariate (Fig. 6 D) and multivariate (Fig. 6 E) regression analyses were illustrated using forest plots, and detailed statistical parameters are provided in Table 4 . Our analysis revealed that the RiskScore demonstrated remarkable statistical significance (p < 0.001) in predicting patient outcomes. Table 4 Results of Cox Analysis Characteristics Total(N) Univariate analysis Multivariate analysis HR (95% CI) P value HR(95% CI) P value Age 176 1.025 (1.003–1.047) 0.025 1.025 (1.004–1.046) 0.021 Gender 176 FEMALE 79 Reference MALE 97 0.788 (0.522–1.190) 0.257 Stage_M 176 M0 79 Reference M1&MX 97 0.851 (0.562–1.287) 0.444 Stage_N 176 N0 49 Reference Reference N1&NX 127 1.991 (1.188–3.337) 0.009 1.727 (1.019–2.929) 0.043 Stage_T 176 T1&T2 30 Reference Reference T3&T4&TX 146 1.781 (0.946–3.356) 0.074 1.176 (0.616–2.245) 0.624 Risk.Score 176 5.158 (2.685–9.910) < 0.001 5.182 (2.668–10.062) 1 indicates that the variable is a risk factor, while an HR < 1 suggests that it is a protective factor. Variables with a univariate p value < 0.1 were included in the analysis The predictive accuracy of our risk stratification model was systematically evaluated through calibration analyses conducted at annual intervals (1-, 2-, and 3-year follow-up periods), with the resulting calibration plots presented in Figs. 6 F through 6 H. The results indicated that the model exhibited optimal predictive accuracy for clinical outcomes at the 1-year interval. Next, we used DCA at the same time intervals to assess the clinical utility (Figs. 6 I–K). Our evaluation revealed that the LASSO regression model exhibited progressively enhanced predictive accuracy for longer-term outcomes, with the highest performance observed for 3-year predictions, followed by 2-year and then 1-year forecasts. To investigate the expression of the model genes between groups, we used a grouped comparison chart to illustrate the expression analysis results (Fig. 6 L). Statistical analysis revealed markedly distinct expression patterns of the five signature genes between high- and low-risk cohorts, with statistical significance (p < 0.001). Subsequently, ROC curve analysis was performed to evaluate the predictive performance of selected model genes The results revealed that LDHA expression displayed excellent discriminative capacity (AUC > 0.9, Fig. 6 N), while IL18, EREG and SOCS2 expressions exhibited intermediate predictive accuracy (0.7 ≤ AUC < 0.9, Figs. 6 M and 6 P). In contrast, SPP1 showed relatively limited discriminatory power (0.5 < AUC < 0.7, Fig. 6 Q) in distinguishing between high- and low-risk patient subgroups. Furthermore, we quantified the stromal, immune, and ESTIMATE scores based on the transcriptomic profiles from the TCGA-PAAD cohort. Comparative analysis between high- and low-risk patient subgroups revealed statistically significant disparities in stromal compartment scores (p < 0.01), as illustrated in Fig. 6 O. These findings demonstrate distinct tumor microenvironment characteristics between the prognostic groups. The immunological and ESTIMATE scoring systems exhibited highly significant differences (p < 0.001) when comparing the two cohorts, with notably elevated values observed in patients classified as low-risk relative to their high-risk counterparts. 4 Discussion Subtype-specific molecular traits can influence treatment responses and clinical outcomes, suggesting that a deeper understanding of these variations could improve therapeutic efficacy [ 24 ]. PAAD represents a highly heterogeneous malignancy characterized by the presence of distinct molecular subtypes. These subtypes exhibit notable differences in prognosis and treatment response [ 25 ]. Statistical analyses reveal that pancreatic adenocarcinoma (PAAD) patients exhibit a dismal prognosis, with fewer than 15% surviving beyond five years post-diagnosis, largely due to the development of drug resistance. Although initial treatments may be effective, resistance often develops rapidly, significantly diminishing therapeutic success [ 26 ]. Within the TME, macrophages serve as pivotal mediators that significantly influence both neoplastic advancement and therapeutic resistance. These cell populations demonstrate significant adaptability, enabling their differentiation into specialized functional subtypes, notably the conventionally activated M1 and alternatively stimulated M2 phenotypes. M1-polarized macrophages are primarily associated with pro-inflammatory responses and exhibit anti-tumor properties, while M2-polarized macrophages tend to promote immunosuppressive functions and facilitate tumor progression [ 27 ]. The involvement of macrophage polarization in mediating drug resistance in PAAD has received increasing attention [ 28 ]. Recent studies have demonstrated that modulating macrophage polarization can affect cancer cell sensitivity to various therapies [ 29 ]. Therefore, macrophage polarization may be linked to drug resistance in PAAD. We initially identified 36 DMRGs from public databases and based on these genes, we distinguished two PAAD subtypes: A (cluster 1) and B (cluster 2). These subtypes exhibit significant clinical differences, with subtype A showing a considerably better prognosis than subtype B (Fig. 1 ). GSEA revealed that PAAD samples were enriched in immune responses and cell signaling pathways, emphasizing the critical role of the TME in cancer progression (Fig. 2 ). Specifically, pathways such as Fceri signaling and FCGR3A-mediated IL-10 production highlight the complex interactions between tumor cells and immune components. Within the TME, immune complexes formed by tumor antigens and IgG antibodies bind to FCGR3A on TAMs, activating the Syk/PI3K/Akt or nuclear factor kappa-beta pathways and promoting IL-10 gene transcription [ 30 ]. Secreted IL-10 can upregulate M2 markers, thus promoting TAM polarization towards an M2 phenotype associated with tumor progression [ 31 ]. Animal models have demonstrated that knocking out FCGR3A or blocking IL-10 reduces M2-type TAM infiltration, inhibiting tumor growth and metastasis [ 32 ]. Therefore, targeting this pathway holds promise for reversing the tumor-promoting effects of TAMs and for developing new strategies for tumor immunotherapy. However, further investigation is needed to elucidate the mechanistic differences across tumor contexts and optimize precision intervention strategies. The identification of individuals with a higher likelihood of positive response to immunotherapeutic interventions is crucial for improving clinical outcomes [ 33 ]. Notably, a substantial difference in TIDE immunotherapy scores existed between subtypes A and B, indicating that subtype B may be more responsive to immune checkpoint inhibitors than subtype A (Fig. 3 ). Immunological profiling using CIBERSORT demonstrated significant intercellular associations within PAAD subtype A, where a marked inverse relationship was observed between monocyte populations and M0 macrophage subsets (r = -0.418, p = 0.032). This analysis highlighted the dynamic interplay between distinct immune cell lineages in PAAD microenvironment. Notably, subtype B exhibited pronounced inverse relationships among immune cell populations, particularly between monocytes and M0 macrophages which displayed the most substantial negative correlation (r = -0.55, p-value = 0.047). These immunological profiling results provide crucial insights for designing precision immunotherapeutic interventions customized to patients with distinct characteristics, ultimately leading to optimized personalized treatment plans and a reduced patient burden. To further validate the prognostic significance of DEGs across the PAAD subtypes, we performed LASSO regression analysis using a panel of 19 DMRDEGs and developed a corresponding risk model (Fig. 5 ). This model could predict PAAD survival rates, paving the way for tailored immunotherapy for patients with PAAD who are susceptible to drug resistance. The model comprised five genes: IL18, EREG, LDHA, SOCS2, and SPP1. CIBERSORT analysis revealed that in subtype A, SOCS2 expression was negatively related to the upregulation of M0 macrophages (r = -0.458, p < 0.05), suggesting that SOCS2 negatively regulates M0 macrophage aggregation. SOCS2, a member of the SOCS family, regulates signaling pathways such as JAK-STAT, and is involved in tumorigenesis, development, and drug resistance [ 34 ]. Mechanistically, SOCS2 hinders tumor cell survival and invasion by inhibiting the JAK2/STAT5 pathway [ 35 ]. In the immune microenvironment, SOCS2 decreases M2-type TAM polarization and immunosuppressive factor secretion by inhibiting IL-6/JAK2/STAT3 signaling [ 36 ]. In malignancies exhibiting reduced SOCS2 levels, particularly in non-small cell lung carcinoma cases, experimental restoration of SOCS2 expression has been demonstrated to suppress metastatic potential and reverse chemotherapeutic resistance [ 37 ]. Similarly, in PAAD, SOCS2 may contribute to the resistance to chemotherapy and targeted therapy via related mechanisms. Restoring SOCS2 function or targeting its downstream pathways may emerge as novel strategies for overcoming drug resistance, pending further translational research. SOCS2 mRNA and protein expression is significantly downregulated in PAAD cells compared to normal pancreatic cells [ 38 ]. Nevertheless, the functional significance of SOCS2 in mediating chemoresistance mechanisms within PAAD has not been fully elucidated. In future, we aim to investigate the dynamic regulatory network of SOCS2 in PAAD drug resistance using single-cell sequencing and organoid models. This investigation presents several limitations that warrant consideration. The modest cohort size may constrain the broader applicability of the results. Furthermore, the lack of experimental validation in laboratory settings limits the confirmation of proposed biomarkers and their mechanistic involvement in pancreatic adenocarcinoma. The integration of diverse datasets also raises the possibility of batch-related variations, which could influence the consistency and clarity of the observed outcomes. 5 Conclusions In conclusion, our analysis categorized PAAD into two distinct subgroups, designated as Subtype A (1) and Subtype B (2), based on DMRGs. And we also successfully identified 34 DMRDEGs that are crucial for understanding the TME in patients. A prognostic risk model developed from these genes exhibits considerable promise for forecasting patient survival outcomes. Future research could pay attention on clarifying the regulatory mechanisms of macrophage polarization and validating these findings in larger clinically relevant cohorts. Finally, this study lays the groundwork for the development of novel therapeutic targets and strategies to improve PAAD treatment outcomes. Abbreviations PAAD, pancreatic adenocarcinoma TCGA, The Cancer Genome Atlas GEO, Gene Expression Omnibus DMRGs, drug resistant and macrophage polarization-related genes DEGs, differentially expressed genes DMRDEGs, drug resistance- and macrophage polarization-related differentially expressed genes TME, tumor microenvironment TAM, tumor-associated macrophage FPKM, fragments per kilobase per million DRGs, drug resistance-related genes MRGs, macrophage polarization-related genes OS, overall survival KM, Kaplan–Meier SM, somatic mutations ICGs, immune checkpoint genes CNV, copy number variations TIDE, Tumor Immune Dysfunction and Exclusion GSEA, gene set enrichment analysis MSigDB, Molecular Signatures Database FDR, false discovery rate BH, Benjamini-Hochberg LASSO, Least Absolute Shrinkage and Selection Operator ROC, receiver operating characteristic AUC, area under the curve DCA, decision curve analysis PPI, protein–protein interaction TF, transcription factor miRNA, microRNA mRNA, messenger RNA RBP, RNA-binding protein IL, interleukin NK, natural killer SNP, single nucleotide polymorphism SNV, single nucleotide variant CDF, empirical cumulative distribution function PCA, principal component analysis Declarations Funding Partial financial support was received from the Youth Development Program of the First Affiliated Hospital of Naval Medical University (No. 2021JCQN04) and General Program Incubation Program of Naval Medical University (No. 2023MS017). Author Contributions Zixin Liu, Ruijiao Kong and Zichen Liu contributed equally to this work. Zixin Liu, Ruijiao Kong, Yin Jia and Gang Jin were involved in the study design; Ruijiao Kong and Zichen Liu collected and processed the data. Zixin Liu, Ruijiao Kong and Zichen Liu analyzed and interpreted the data. Zixin Liu, Ruijiao Kong and Yin Jia contributed to the analysis methods. Zixin Liu, Ruijiao Kong, Zichen Liu and Gang Jin contributed to the writing of the manuscript. All the authors approved the final manuscript. Conflicts of Interests The authors have no relevant financial or non-financial interests to disclose. Ethics approval and consent to participate Since the data adopted in this study were all publicly available data from the TCGA, GEO, GeneCards and DRESIS database, all data related studies were approved by their respective ethical review committees and received written informed consent from patients. Therefore, this study does not need additional ethics approval. Consent to publish Not applicable. Data Availability Statement Data are available on reasonable request. All data relevant to the study are included in the article or uploaded as supplemental information.TCGA_PAAD is downloaded from The Cancer Genome Atlas (TCGA) database (https://www.cancer.gov/ccg/research/genome-sequencing/tcga). Microarray data sets (GSE28735 and GSE71729) are downloaded from the Gene Expression Integrated Database (GEO) (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE28735 and https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE71729). References Yueting L, Xin S, Shuai L et al. Development and validation of a predictive model based upon extracellular vesicle-derived transposable elements for non-invasive detection of pancreatic adenocarcinoma. 2025;13(1). https://doi.org/10.1186/s40364-025-00770-6 Xiaolan H, Zhengyang X, Ruiping R et al. A novel sphingolipid metabolism-related long noncoding RNA signature predicts the prognosis, immune landscape and therapeutic response in pancreatic adenocarcinoma. 2024;10(1). https://doi.org/10.1016/j.heliyon.2023.e23659 Liuyuan X, Xin G, Dong L et al. S100A13-driven interaction between pancreatic adenocarcinoma cells and cancer-associated fibroblasts promotes tumor progression through calcium signaling. 2025;23(1). https://doi.org/10.1186/s12964-025-02049-7 Lan THP, Yating C, Yohei F et al. AXL promotes inflammatory breast cancer progression by regulating immunosuppressive macrophage polarization. 2025;27(1). https://doi.org/10.1186/s13058-025-02015-8 Christopher JH, Corbin P, Ilya K et al. Macrophage-Released Pyrimidines Inhibit Gemcitabine Therapy in Pancreatic Cancer. 2019;29(6). https://doi.org/10.1016/j.cmet.2019.02.001 Hibiki U, Kunitoshi S, Toshiaki T et al. ADAR1-high tumor-associated macrophages induce drug resistance and are therapeutic targets in colorectal cancer. 2025;24(1). https://doi.org/10.1186/s12943-025-02312-y Barrett T, Wilhite SE, Ledoux P, et al. NCBI GEO: archive for functional genomics data sets–update. Nucleic Acids Res. 2013. https://doi.org/10.1093/nar/gks1193 . 41(Database issue):D991-5. Stelzer G, Rosen N, Plaschkes I, et al. Curr protocols Bioinf. 2016;54. https://doi.org/10.1002/cpbi.5 . The GeneCards Suite: From Gene Data Mining to Disease Genome Sequence Analyses1.30.1-1.3. Sun X, Zhang Y, Li H, et al. DRESIS: the first comprehensive landscape of drug resistance information. Nucleic Acids Res. 2023;51(D1):D1263–75. https://doi.org/10.1093/nar/gkac812 . Yin F, Liu X, Li D, Wang Q, Zhang W, Li L. Tumor suppressor genes associated with drug resistance in ovarian cancer (review). Oncol Rep. 2013;30(1):3–10. https://doi.org/10.3892/or.2013.2446 . Yu H, Zhang W, Xu XR, Chen S. Drug resistance related genes in lung adenocarcinoma predict patient prognosis and influence the tumor microenvironment. Sci Rep. 2023;13(1):9682. https://doi.org/10.1038/s41598-023-35743-y . Wang J, Li X, Wang K, et al. CLEC7A regulates M2 macrophages to suppress the immune microenvironment and implies poorer prognosis of glioma. Front Immunol. 2024;15:1361351. https://doi.org/10.3389/fimmu.2024.1361351 . Li S, Zhang N, Yang Y, Liu T. Transcriptionally activates CCL28 expression to inhibit M2 polarization of macrophages and prevent immune escape in colorectal cancer cells. Translational Oncol. 2024;40:101842. https://doi.org/10.1016/j.tranon.2023.101842 . Yang Z, Wang J, Pi J, et al. Identification and Validation of Genes Related to Macrophage Polarization and Cell Death Modes Under Mycobacterium tuberculosis Infection. J Inflamm Res. 2024;17:1397–411. https://doi.org/10.2147/jir.S448372 . Hu P, Li B, Yin Z, et al. Multi-omics characterization of macrophage polarization-related features in osteoarthritis based on a machine learning computational framework. Heliyon. 2024;10(9):e30335. https://doi.org/10.1016/j.heliyon.2024.e30335 . Shi Z, Shen J, Qiu J, Zhao Q, Hua K, Wang H. CXCL10 potentiates immune checkpoint blockade therapy in homologous recombination-deficient tumors. Theranostics. 2021;11(15):7175–87. https://doi.org/10.7150/thno.59056 . Bao Y, Jiang A, Dong K, et al. DDX39 as a predictor of clinical prognosis and immune checkpoint therapy efficacy in patients with clear cell renal cell carcinoma. Int J Biol Sci. 2021;17(12):3158–72. https://doi.org/10.7150/ijbs.62553 . Jiang P, Gu S, Pan D, et al. Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nat Med. 2018;24(10):1550–8. https://doi.org/10.1038/s41591-018-0136-1 . Liberzon A, Subramanian A, Pinchback R, Thorvaldsdóttir H, Tamayo P, Mesirov JP. Molecular signatures database (MSigDB) 3.0. Bioinformatics (Oxford. England). 2011;27(12):1739–40. https://doi.org/10.1093/bioinformatics/btr260 . Szklarczyk D, Gable AL, Lyon D, et al. STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res. 2019;47(D1):D607–13. https://doi.org/10.1093/nar/gky1131 . Franz M, Rodriguez H, Lopes C, et al. GeneMANIA update 2018. Nucleic Acids Res. 2018;46(W1):W60–4. https://doi.org/10.1093/nar/gky311 . Zhou KR, Liu S, Sun WJ, et al. ChIPBase v2.0: decoding transcriptional regulatory networks of non-coding RNAs and protein-coding genes from ChIP-seq data. Nucleic Acids Res. 2017;45(D1):D43–50. https://doi.org/10.1093/nar/gkw965 . Li JH, Liu S, Zhou H, Qu LH, Yang JH. Nucleic Acids Res. 2014;42(Database issue):D92–7. https://doi.org/10.1093/nar/gkt1248 . starBase v2.0: decoding miRNA-ceRNA, miRNA-ncRNA and protein-RNA interaction networks from large-scale CLIP-Seq data. Deepak R et al. In Hwan P, Timothy V P,. Syk Inhibition Reprograms Tumor-Associated Macrophages and Overcomes Gemcitabine-Induced Immunosuppression in Pancreatic Ductal Adenocarcinoma. 2023;83(16). https://doi.org/10.1158/0008-5472.Can-22-3645 Ben G, Olga K, Andrey K et al. Transcriptomic-Based Microenvironment Classification Reveals Precision Medicine Strategies for Pancreatic Ductal Adenocarcinoma. 2024;166(5). https://doi.org/10.1053/j.gastro.2024.01.028 Rebecca LS, Angela NG, Ahmedin JJCCJC, Cancer statistics. 2024. 2024;74(1). https://doi.org/10.3322/caac.21820 Daan JK, Leila AJC. Macrophages at the interface of the co-evolving cancer ecosystem. 2023;186(8). https://doi.org/10.1016/j.cell.2023.02.020 Núria V-B, Neus M-B, Pablo GF, Pilar NJE. Hallmarks of pancreatic cancer: spotlight on TAM receptors. 2024;107(0). https://doi.org/10.1016/j.ebiom.2024.105278 Hong CE, Lyu SY. Modulation of Breast Cancer Cell Apoptosis and Macrophage Polarization by Mistletoe Lectin in 2D and 3D Models. Int J Mol Sci. 2024;25(15). https://doi.org/10.3390/ijms25158459 . Hye-Soo P, Yong Woo B, Yeo-Jin S, Hwa-Jung KJC. Mycobacterium avium subsp. paratuberculosis MAP1889c Protein Induces Maturation of Dendritic Cells and Drives Th2-biased Immune Responses. 2020;9(4). https://doi.org/10.3390/cells9040944 Qing L, Chaogang Y, Shuyi W et al. Wnt5a-induced M2 polarization of tumor-associated macrophages via IL-10 promotes colorectal cancer progression. 2020;18(1). https://doi.org/10.1186/s12964-020-00557-2 Jennifer S, Stefan C, Feda A et al. IL-10R inhibition reprograms tumor-associated macrophages and reverses drug resistance in multiple myeloma. 2024;38(11). https://doi.org/10.1038/s41375-024-023.91-8 Thomas FS, Ammar AJ, Atsushi O, et al. Pancreat cancer. 2025;405(10485). https://doi.org/10.1016/s0140-6736(25)00261-2 . Eun-Sook P, Mehrangiz D, Ah Ram L, Bo Bae P, Kyun-Hwan KJC. Dysregulation of Liver Regeneration by Hepatitis B Virus Infection: Impact on Development of Hepatocellular Carcinoma. 2022;14(15). https://doi.org/10.3390/cancers14153566 Yong L, Xiaolong X, Guoyou Z et al. SOCS2 inhibits hepatoblastoma metastasis via downregulation of the JAK2/STAT5 signal pathway. 2023;13(1). https://doi.org/10.1038/s41598-023-48591-7 Dan L, Yandan H, Min W, Bin C, Yan LJCCHTS. Overexpression of SOCS2 Inhibits EMT and M2 Macrophage Polarization in Cervical Cancer via IL-6/JAK2/STAT3 Pathway. 2023;27(7). https://doi.org/10.2174/1386207326666230818092532 Yuan Z, Jinheng Q, Yanneng F et al. Inhibition of DDR1 promotes ferroptosis and overcomes gefitinib resistance in non-small cell lung cancer. 2024;1870(7). https://doi.org/10.1016/j.bbadis.2024.167447 Mingjie H, Yi C, Zhiping Y, Lixia Z, Huamei LJABP. Upregulation of SOCS2 causes mitochondrial dysfunction and promotes ferroptosis in pancreatic cancer cells. 2023;70(1). https://doi.org/10.18388/abp.2020_6383 Additional Declarations No competing interests reported. Supplementary Files Supplementarymaterial1Fig.S1.tif Supplementary material Supplementary material 1 - Fig. S1 Batch effect removal from GSE28735 and GSE71729. A. Box plot showing the distribution of PAAD datasets before batch effect removal. B. Box plot illustrating the distribution of the integrated GEO datasets after batch effect correction. C. PCA plot of PAAD datasets prior to batch effect removal. D. PCA plot of the integrated GEO datasets following batch effect adjustment. PCA, principal component analysis. The yellow color represents the PAAD dataset GSE28735, whereas the purple color denotes the PAAD dataset GSE71729 Abbreviations: PAAD, pancreatic adenocarcinoma; GEO, Gene Expression Omnibus; PCA, principal component analysis Supplementarymaterial2Fig.S2.tif Supplementary material 2 - Fig. S2 Delta plot of the consensus clustering analysis Supplementarymaterial3Fig.S3.tif Supplementary material 3 - Fig. S3 mRNA–miRNA regulatory network of model genes Supplementarymaterial4TableS1DMRGs.xls Supplementary material 4 - Table S1. Genes related to drug resistance and macrophage polarization (DMRGs) Supplementarymaterial5TableS2ICGs.xls Supplementary material 5 - Table S2. Genes related to immune checkpoint genes (ICGs) from the PubMed literature. Supplementarymaterial6TableS3DMRDEGs.xls Supplementary material 6 - Table S3. Drug resistance- and macrophage polarization-related differentially expressed genes (DMRDEGs) Supplementarymaterial7TableS4mRNATF.xls Supplementary material 7 - Table S4. Details on mRNA–transcription factor (TF) regulatory network Supplementarymaterial8TableS5mRNAmiRNA.xls Supplementary material 8 - Table S5. Details on mRNA-miRNA regulatory network Supplementarymaterial9TableS6mRNARBP.xls Supplementary material 9 - Table S6. Details on mRNA- RNA-binding protein (RBP) regulatory network Cite Share Download PDF Status: Posted Version 1 posted 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-7991235","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":559925487,"identity":"c9c02b22-ed26-43fa-a5db-b9546f7023bd","order_by":0,"name":"Zixin Liu","email":"","orcid":"","institution":"Changhai Hospital, Navy Military Medical University","correspondingAuthor":false,"prefix":"","firstName":"Zixin","middleName":"","lastName":"Liu","suffix":""},{"id":559925489,"identity":"de414d84-1fae-4f73-9d10-5aa285fa69a7","order_by":1,"name":"Ruijiao Kong","email":"","orcid":"","institution":"Changhai Hospital, Naval Medical University","correspondingAuthor":false,"prefix":"","firstName":"Ruijiao","middleName":"","lastName":"Kong","suffix":""},{"id":559925490,"identity":"26886017-fe03-4506-8e20-c703baa2c5ae","order_by":2,"name":"Zichen Liu","email":"","orcid":"","institution":"Wenzhou Medical University","correspondingAuthor":false,"prefix":"","firstName":"Zichen","middleName":"","lastName":"Liu","suffix":""},{"id":559925491,"identity":"e794f196-b7dd-43b3-a942-6e48316ef525","order_by":3,"name":"Yin Jia","email":"","orcid":"","institution":"Changhai Hospital, Naval Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yin","middleName":"","lastName":"Jia","suffix":""},{"id":559925492,"identity":"0deb8985-1329-42c4-82c1-f1533a868545","order_by":4,"name":"Gang Jin","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA0ElEQVRIiWNgGAWjYBACxmYILQflMxOvxZh4LTCQ2EC0FuZ23sMvfu6oTZ8/7XSaBEOFdWID+9kDBBzGl2bZe+Z47obbudskGM6kJzbw5CUQ0MJjZsDbdix3gzRQC2Pb4cQGCR4DgloM/7YdS5efDdLyjzgtxo9522oSGEAOY2wg0hZm2bYDhkC/bLZIOJZu3MaTg1+LYf8Z449v2+rkgQ7beONDjbVsP/sZAloaGNgkGBgOQ3gJQMyGVz0QyAOj5gMDQx0hdaNgFIyCUTCSAQAkZUQnw/d11gAAAABJRU5ErkJggg==","orcid":"","institution":"Changhai Hospital, Navy Military Medical University","correspondingAuthor":true,"prefix":"","firstName":"Gang","middleName":"","lastName":"Jin","suffix":""}],"badges":[],"createdAt":"2025-10-30 16:08:25","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7991235/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7991235/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":98777239,"identity":"f32d6be1-1632-4885-9714-032870be9cd6","added_by":"auto","created_at":"2025-12-22 12:26:11","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3782797,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript11.12.docx","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/15de89ee30042be82c7727d6.docx"},{"id":98777889,"identity":"53d8435c-0a6a-42c9-a1d3-0e9cbf903d22","added_by":"auto","created_at":"2025-12-22 12:28:37","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":7526,"visible":true,"origin":"","legend":"","description":"","filename":"94ba27e78bb94101993fc02c35afca53.json","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/c49368ff0576c2f198e666d3.json"},{"id":98747146,"identity":"11d7a84b-f6f0-4b77-8e0a-0d0b3bbfc673","added_by":"auto","created_at":"2025-12-22 08:50:32","extension":"tif","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3210460,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial1Fig.S1.tif","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/e27cc14d4403d425bde4ee74.tif"},{"id":98747156,"identity":"bc743cf8-0c76-41b9-b48f-d8559eebb47c","added_by":"auto","created_at":"2025-12-22 08:50:33","extension":"tif","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":278412,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial2Fig.S2.tif","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/11085efae1c078d2314e1688.tif"},{"id":98777322,"identity":"0e33112e-6ddd-4062-9ba1-87fdb0becd7b","added_by":"auto","created_at":"2025-12-22 12:26:32","extension":"tif","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":411316,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial3Fig.S3.tif","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/4ed80e9894fecec93f7bfd8e.tif"},{"id":98777160,"identity":"d6b96179-d5f9-4406-a643-53e173395d38","added_by":"auto","created_at":"2025-12-22 12:25:31","extension":"xls","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":25088,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial4TableS1DMRGs.xls","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/e1f2c1084b41a37e12096ebd.xls"},{"id":98779825,"identity":"0895444c-6c04-436d-9e7f-54525ca9e34d","added_by":"auto","created_at":"2025-12-22 12:30:47","extension":"xls","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":20992,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial5TableS2ICGs.xls","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/075838c1e9dc8666a7c18ca7.xls"},{"id":98747163,"identity":"09ccf6e6-92db-434d-b751-f62f06ff7517","added_by":"auto","created_at":"2025-12-22 08:50:33","extension":"xls","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":21504,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial6TableS3DMRDEGs.xls","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/bd42656a3f96f178418b799d.xls"},{"id":98747185,"identity":"24221818-b06b-4c0a-834e-a5e500a4a852","added_by":"auto","created_at":"2025-12-22 08:50:33","extension":"xls","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":22528,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial7TableS4mRNATF.xls","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/f879c8670674a180a17202a4.xls"},{"id":98777167,"identity":"46d2cdd8-28ea-46d8-b220-d72a3d5c2c4b","added_by":"auto","created_at":"2025-12-22 12:25:32","extension":"xls","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":23040,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial8TableS5mRNAmiRNA.xls","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/b49d113ff8c2a1d77bb4fe33.xls"},{"id":98778308,"identity":"a1d09fa0-c285-4d2e-ae38-8dc2cf62e6fd","added_by":"auto","created_at":"2025-12-22 12:29:09","extension":"xls","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":22528,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial9TableS6mRNARBP.xls","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/f716c3c808d253e037ee3e6d.xls"},{"id":98747169,"identity":"eb84cc3b-d50a-4a78-9963-cad9c5d1367b","added_by":"auto","created_at":"2025-12-22 08:50:33","extension":"xml","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":150173,"visible":true,"origin":"","legend":"","description":"","filename":"94ba27e78bb94101993fc02c35afca531enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/63d9bbf8cf7aae435793f1c5.xml"},{"id":98777418,"identity":"03172844-fd5e-40c2-8ca0-d3b8d19fe2bd","added_by":"auto","created_at":"2025-12-22 12:27:03","extension":"jpeg","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":757858,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/2e57848191774c15b60ac986.jpeg"},{"id":98747171,"identity":"18f47af6-1bac-4872-90e8-8769f47e7572","added_by":"auto","created_at":"2025-12-22 08:50:33","extension":"jpeg","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":475196,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/3ead7f564c29b96fada787cc.jpeg"},{"id":98777141,"identity":"fc059105-8bf0-410b-8b75-a0cb55052c41","added_by":"auto","created_at":"2025-12-22 12:25:29","extension":"jpeg","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":584796,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/8647d852b37bb51c54fdc989.jpeg"},{"id":98747170,"identity":"4ec3e4a4-82e3-4f48-8225-d873d4332272","added_by":"auto","created_at":"2025-12-22 08:50:33","extension":"jpeg","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":616346,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/fdc0b60739f0545340bec4ec.jpeg"},{"id":98777096,"identity":"506b50f0-4feb-412e-b973-a0736df5e5b6","added_by":"auto","created_at":"2025-12-22 12:25:21","extension":"jpeg","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1077524,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/1731ce62e2e7370b855d46bd.jpeg"},{"id":98779204,"identity":"607521e4-ce8b-42f7-8348-e1b13ec860b3","added_by":"auto","created_at":"2025-12-22 12:30:04","extension":"jpeg","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":532286,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/19999ebcfbef600d8c037654.jpeg"},{"id":98747166,"identity":"e9cc7e80-ec1d-4064-a2d0-d27322d95ffe","added_by":"auto","created_at":"2025-12-22 08:50:33","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":92026,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/db0cc4b9b7e6e18accdd130c.png"},{"id":98747167,"identity":"7b1e4130-97ca-476f-9e3f-0b5648ffbf56","added_by":"auto","created_at":"2025-12-22 08:50:33","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":69010,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/3f245eee0ba48ac22ead92df.png"},{"id":98747172,"identity":"9da5bd71-6c03-402f-8800-2d87949a9f3d","added_by":"auto","created_at":"2025-12-22 08:50:33","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":78696,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/ae6b433b95a547d38d11fc5b.png"},{"id":98747184,"identity":"1e744ed2-262f-4f8e-b291-a4865b4f497f","added_by":"auto","created_at":"2025-12-22 08:50:33","extension":"png","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":49568,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/8c043f8d8ff11869644153fb.png"},{"id":98747175,"identity":"6ad6c804-32cf-472d-bd78-d1df98c986be","added_by":"auto","created_at":"2025-12-22 08:50:33","extension":"png","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":140958,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/327a04bc3e8367995daa156c.png"},{"id":98747173,"identity":"80353225-155f-41d8-9161-8ef38ccd6b73","added_by":"auto","created_at":"2025-12-22 08:50:33","extension":"png","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":80038,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/1bead0e01f90960d612216ec.png"},{"id":98747174,"identity":"dfb58bc8-8a03-4e80-8821-1ae93f800487","added_by":"auto","created_at":"2025-12-22 08:50:33","extension":"xml","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":148842,"visible":true,"origin":"","legend":"","description":"","filename":"94ba27e78bb94101993fc02c35afca531structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/19af8c0a2a6b89f580e723da.xml"},{"id":98747186,"identity":"e18d4a18-1bd5-4aad-b48a-2a2a3dd06db1","added_by":"auto","created_at":"2025-12-22 08:50:33","extension":"html","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":164519,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/387f288b4db7ba3a4497e246.html"},{"id":98747142,"identity":"9f8f1183-5410-46a0-a24e-75f6d382f0fe","added_by":"auto","created_at":"2025-12-22 08:50:32","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":92026,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIdentification of two subtypes of PAAD based on drug resistance\u003c/strong\u003e-\u003cstrong\u003e and macrophage polarization-related genes (DMRGs).\u003c/strong\u003e \u003cstrong\u003e(A)\u003c/strong\u003eFlow chart for the comprehensive analysis of DMRDEGs. \u003cstrong\u003e(B) \u003c/strong\u003eForest Plot of 36 DMRGs in the univariate Cox regression model. \u003cstrong\u003e(C) \u003c/strong\u003eConsensus clustering result of PAAD samples from the TCGA-PAAD cohort. \u003cstrong\u003e(D)\u003c/strong\u003e CDF plot of the consensus clustering analysis. \u003cstrong\u003e(E) \u003c/strong\u003e3D t-SNE clustering plot of the two PADD subtypes. \u003cstrong\u003e(F)\u003c/strong\u003eVolcano plot of DEG analysis across different disease subtypes in the TCGA-PAAD cohort. \u003cstrong\u003e(G)\u003c/strong\u003e Venn diagram showing DEGs and DMRGs. \u003cstrong\u003e(H)\u003c/strong\u003e Heatmap of the top 20 DMRDEGs ranked by absolute logFC value.\u003cstrong\u003e (I) \u003c/strong\u003eChromosome localization map of DMRDEGs. Purple represents subtype A (Cluster 1), and yellow represents subtype B (Cluster 2). In the heatmap, red indicates high expression, and blue indicates low expression. “ns” represents p ≥ 0.05, no statistical significance; “*” represents p \u0026lt; 0.05, statistical significance; “**” represents p \u0026lt; 0.01, high statistical significance; and “***” represents p \u0026lt; 0.001, extreme statistical significance\u003c/p\u003e","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/3401017a7bdcd796477ab5a9.png"},{"id":98747143,"identity":"31795bbd-b707-486f-87a4-db58d554f0b5","added_by":"auto","created_at":"2025-12-22 08:50:32","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":69010,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eClinical characteristics of the two PAAD subtypes. (A) \u003c/strong\u003eFour biological functions identified using a mountain plot for the GSEA of the PAAD dataset. \u003cstrong\u003e(B)–(E)\u003c/strong\u003e depict the GSEA results, demonstrating significant enrichment of all genes in Fc epsilon receptor (Fceri) signaling \u003cstrong\u003e(B)\u003c/strong\u003e, FCGR3A-mediated IL-10 synthesis \u003cstrong\u003e(C)\u003c/strong\u003e, Fceri-mediated Ca2+ mobilization \u003cstrong\u003e(D)\u003c/strong\u003e, and LAT2/NTAL pathway involvement in calcium mobilization \u003cstrong\u003e(E)\u003c/strong\u003e. Selection criteria for GSEA are set based on adj.p \u0026lt; 0.05 and an FDR value (q value) \u0026lt;0.25, using the being Benjamini–Hochberg (BH) adj.p adjustment method. \u003cstrong\u003e(F)\u003c/strong\u003e KM curves for prognosis between subtypes A and B in the TCGA-PAAD dataset compared with the OS of PAAD samples. \u003cstrong\u003e(G)\u003c/strong\u003e Proportions of patients in subtype A (Cluster 1) and subtype B (Cluster 2) in the TCGA-PAAD dataset according to age, gender, M stage, N stag, and T stage. Purple represents subtype A, whereas yellow represents subtype B. “ns” represents p ≥ 0.05, no statistical significance; “*” represents p \u0026lt; 0.05, statistical significance; “**” represents p \u0026lt; 0.01, high statistical significance; and “***” represents p \u0026lt; 0.001, extreme statistical significance\u003c/p\u003e","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/92e110a1e1fc9fc06f880ee9.png"},{"id":98778890,"identity":"ac91ca9e-5aff-4f7e-960c-fe2a427d816e","added_by":"auto","created_at":"2025-12-22 12:29:47","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":78696,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDifferences in the immune environments between the two PAAD subtypes. (A)\u003c/strong\u003e Comparison of immune cell grouping in subtype A (Cluster 1) and subtype B (Cluster 2) of PAAD samples. \u003cstrong\u003e(B), (C)\u003c/strong\u003e Correlation heatmaps of immune cells in Cluster 1 \u003cstrong\u003e(B) \u003c/strong\u003eand Cluster 2 \u003cstrong\u003e(C)\u003c/strong\u003e of PAAD samples.\u003cstrong\u003e (D), (E)\u003c/strong\u003e Comparison of immune checkpoint genes \u003cstrong\u003e(D) \u003c/strong\u003eand TIDE scores \u003cstrong\u003e(E)\u003c/strong\u003e between subtype A (Cluster 1) and subtype B (Cluster 2) in the PAAD group of the TCGA-PAAD dataset. The absolute value of the correlation coefficient (r-value) is considered weak or uncorrelated at \u0026lt;0.3, weakly correlated between 0.3 and 0.5, moderately correlated between 0.5 and 0.8, and strongly correlated at \u0026gt;0.8. Purple represents subtype A (cluster 1) and yellow represents subtype B (cluster 2). Red and blue indicate positive and negative correlations, respectively. The shaded area represents the strength of the correlation. “ns” represents p ≥ 0.05, no statistical significance; “*” represents p \u0026lt; 0.05, statistical significance; “**” represents p \u0026lt; 0.01, high statistical significance; and “***” represents p \u0026lt; 0.001, extreme statistical significance\u003c/p\u003e","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/135aa7d419cffbb67629c368.png"},{"id":98776848,"identity":"6d8853f1-fe3e-499d-97e1-97f99dd26d9b","added_by":"auto","created_at":"2025-12-22 12:23:43","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":49568,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAnalysis of CNVs and SMs in DMRDEGs based the two PAAD subtypes.\u003c/strong\u003e \u003cstrong\u003e(A), (B) \u003c/strong\u003eExpression of SMs in DMRDEGs within the PAAD group of the TCGA-PAAD dataset.\u003cstrong\u003e (C), (D) \u003c/strong\u003eLollipop plot \u003cstrong\u003e(C)\u003c/strong\u003e and mutation frequency histogram \u003cstrong\u003e(D) \u003c/strong\u003eillustrating CNV types in DMRDEGs. “ns” represents p ≥ 0.05, no statistical significance; “*” represents p \u0026lt; 0.05, statistical significance; “**” represents p \u0026lt; 0.01, high statistical significance; and “***” represents p \u0026lt; 0.001, extreme statistical significance\u003c/p\u003e","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/8100465fd79760376ca54a13.png"},{"id":98747145,"identity":"2f95635d-01a4-442f-b0b4-246e3c513dc3","added_by":"auto","created_at":"2025-12-22 08:50:32","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":140958,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eConstruction of a prognostic risk model for pancreatic cancer with DMRDEGs.\u003c/strong\u003e \u003cstrong\u003e(A) \u003c/strong\u003eForest plot of 19 DMRDEGs in the univariate Cox regression model. \u003cstrong\u003e(B), (C) \u003c/strong\u003eVariable trajectory plot \u003cstrong\u003e(B)\u003c/strong\u003e and the prognostic risk model plot \u003cstrong\u003e(C) \u003c/strong\u003eof the LASSO regression model. \u003cstrong\u003e(D)\u003c/strong\u003e PPI network of model genes established using the STRING database. \u003cstrong\u003e(E)\u003c/strong\u003e GeneMANIA platform to predict the interaction network of functionally similar genes for hub genes related to drug resistance and macrophage polarization. The colors of the connecting lines represent the functionally interconnected relationships.\u003cstrong\u003e (F)\u003c/strong\u003e mRNA–TF regulatory network of model genes.Yellow represents mRNA, blue represents TF, and green represents RBP. \u003cstrong\u003e(G)\u003c/strong\u003e mRNA-RBP regulatory network of model genes.\u003cstrong\u003e (H), (I) \u003c/strong\u003eBubble plots showing the correlation between immune cell infiltration abundance and model genes in Cluster 1 \u003cstrong\u003e(H)\u003c/strong\u003e and Cluster 2 \u003cstrong\u003e(I)\u003c/strong\u003e of PAAD samples. “ns” represents p ≥ 0.05, no statistical significance; “*” represents p \u0026lt; 0.05, statistical significance; “**” represents p \u0026lt; 0.01, high statistical significance; and “***” represents p \u0026lt; 0.001, extreme statistical significance\u003c/p\u003e","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/93ab955b822be3183467afbe.png"},{"id":98779630,"identity":"5450ce9a-9ddc-49eb-a457-3f2fdbacea02","added_by":"auto","created_at":"2025-12-22 12:30:33","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":80038,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eClinical evaluation of the prognostic risk model. (A)\u003c/strong\u003e Time-dependent ROC curve for PAAD samples in the TCGA-PAAD dataset. An AUC of \u0026gt;0.5 indicates a trend toward the promotion of an event by molecular expression, with an AUC closer to 1 signifying better diagnostic performance. \u003cstrong\u003e(B)\u003c/strong\u003e Prognostic KM curve demonstrating the association between high and low RiskScore groups and OS in PAAD samples.\u003cstrong\u003e (C) \u003c/strong\u003eNomogram illustrating the roles of RiskScore and clinical information in both univariate and multivariate Cox regression models.\u003cstrong\u003e (D), (E) \u003c/strong\u003eForest plots displaying RiskScore and clinical information in univariate\u003cstrong\u003e (D) \u003c/strong\u003eand multivariate \u003cstrong\u003e(E) \u003c/strong\u003eCox regression models. \u003cstrong\u003e(F)–(H) \u003c/strong\u003erepresent the 1- \u003cstrong\u003e(F)\u003c/strong\u003e, 2- \u003cstrong\u003e(G)\u003c/strong\u003e, and 3-year \u003cstrong\u003e(H) \u003c/strong\u003ecalibration curves of the prognostic risk model for PAAD. \u003cstrong\u003e(I)–(K)\u003c/strong\u003e DCA for the same prognostic risk model at the 1-year \u003cstrong\u003e(I)\u003c/strong\u003e, 2-year \u003cstrong\u003e(J)\u003c/strong\u003e, and 3-year \u003cstrong\u003e(K)\u003c/strong\u003e time points. \u003cstrong\u003e(L)\u003c/strong\u003e Comparison of the high- and low-risk groups in PAAD samples from the combined GEO datasets, based on model genes. ROC curves for model genes IL-18 and EREG \u003cstrong\u003e(M)\u003c/strong\u003e, LDHA\u003cstrong\u003e (N)\u003c/strong\u003e, SOCS2 \u003cstrong\u003e(P)\u003c/strong\u003e, and SPP1\u003cstrong\u003e (Q)\u003c/strong\u003e in PAAD samples from the combined GEO datasets. \u003cstrong\u003e(O)\u003c/strong\u003e Comparison of immune scores from ESTIMATE analysis between high- and low-risk groups in the TCGA-PAAD dataset. “ns” represents p ≥ 0.05, no statistical significance; “*” represents p \u0026lt; 0.05, statistical significance; “**” represents p \u0026lt; 0.01, high statistical significance; and “***” represents p \u0026lt; 0.001, extreme statistical significance\u003c/p\u003e","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/1f1eca809c7eace7af40dc14.png"},{"id":103266779,"identity":"66c7df6d-601b-40a3-acf3-e6938c7ef150","added_by":"auto","created_at":"2026-02-23 19:55:08","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2169115,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/0161072c-0030-49c8-8444-1f72e5cd8cd0.pdf"},{"id":98747160,"identity":"71d5cf8c-240b-40d7-93aa-4648d9adab40","added_by":"auto","created_at":"2025-12-22 08:50:33","extension":"tif","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":3210460,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSupplementary material\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSupplementary material 1 - \u003cstrong\u003eFig. S1 Batch effect removal from GSE28735 and GSE71729\u003c/strong\u003e. A. Box plot showing the distribution of PAAD datasets before batch effect removal. B. Box plot illustrating the distribution of the integrated GEO datasets after batch effect correction. C. PCA plot of PAAD datasets prior to batch effect removal. D. PCA plot of the integrated GEO datasets following batch effect adjustment. PCA, principal component analysis. The yellow color represents the PAAD dataset GSE28735, whereas the purple color denotes the PAAD dataset GSE71729\u003c/p\u003e\n\u003cp\u003eAbbreviations: PAAD, pancreatic adenocarcinoma; GEO, Gene Expression Omnibus; PCA, principal component analysis\u003c/p\u003e","description":"","filename":"Supplementarymaterial1Fig.S1.tif","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/b53d0bf82b310f51106010c5.tif"},{"id":98778533,"identity":"a1f475a9-aad2-40a5-9dd8-363a5f56ea6c","added_by":"auto","created_at":"2025-12-22 12:29:25","extension":"tif","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":278412,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary material 2 - \u003cstrong\u003eFig. S2 Delta plot of the consensus clustering analysis\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Supplementarymaterial2Fig.S2.tif","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/0d350ab15dbe0d5d878f9a2f.tif"},{"id":98747148,"identity":"0afa3e61-20ce-407b-aa91-3efd6b559801","added_by":"auto","created_at":"2025-12-22 08:50:32","extension":"tif","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":411316,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary material 3 - \u003cstrong\u003eFig. S3 mRNA–miRNA regulatory network of model genes\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Supplementarymaterial3Fig.S3.tif","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/4ac6f19c1de395b586290f79.tif"},{"id":98777059,"identity":"fc57d5b3-b0e2-42a6-85a7-2e0a119df081","added_by":"auto","created_at":"2025-12-22 12:25:18","extension":"xls","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":25088,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary material 4\u003cstrong\u003e - Table S1. Genes related to drug resistance and macrophage polarization (DMRGs)\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Supplementarymaterial4TableS1DMRGs.xls","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/2490d787453d1326e3592a79.xls"},{"id":98777145,"identity":"284c72ba-1b90-4737-9e6d-5515cd3c2ba3","added_by":"auto","created_at":"2025-12-22 12:25:30","extension":"xls","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":20992,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary material 5\u003cstrong\u003e - Table S2. Genes related to immune checkpoint genes (ICGs) from the PubMed literature.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Supplementarymaterial5TableS2ICGs.xls","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/c30fa6dca45d189200739156.xls"},{"id":98778999,"identity":"3ab9070c-e576-4286-913c-b86cf0f20b9d","added_by":"auto","created_at":"2025-12-22 12:29:52","extension":"xls","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":21504,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary material 6 \u003cstrong\u003e- Table S3. Drug resistance- and macrophage polarization-related differentially expressed genes (DMRDEGs)\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Supplementarymaterial6TableS3DMRDEGs.xls","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/7b42a332a70117effd575e1b.xls"},{"id":98747165,"identity":"f5a65248-ae9f-4b84-a704-fed2673ef31e","added_by":"auto","created_at":"2025-12-22 08:50:33","extension":"xls","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":22528,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary material 7\u003cstrong\u003e - Table S4. Details on mRNA–transcription factor (TF) regulatory network\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Supplementarymaterial7TableS4mRNATF.xls","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/8a8ac6a69de6372f82a0978c.xls"},{"id":98747176,"identity":"19c35e67-d4f8-4574-9ccb-8f59fc07fa6f","added_by":"auto","created_at":"2025-12-22 08:50:33","extension":"xls","order_by":8,"title":"","display":"","copyAsset":false,"role":"supplement","size":23040,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary material 8 \u003cstrong\u003e- Table S5. Details on mRNA-miRNA regulatory network\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Supplementarymaterial8TableS5mRNAmiRNA.xls","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/c284d8f4085413022e4e5fd3.xls"},{"id":98777411,"identity":"20dd0b04-5886-4fe7-8320-189f786619ef","added_by":"auto","created_at":"2025-12-22 12:26:59","extension":"xls","order_by":9,"title":"","display":"","copyAsset":false,"role":"supplement","size":22528,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary material 9\u003cstrong\u003e - Table S6. Details on mRNA- RNA-binding protein (RBP) regulatory network\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Supplementarymaterial9TableS6mRNARBP.xls","url":"https://assets-eu.researchsquare.com/files/rs-7991235/v1/7db3e2771489f9c45cbd048d.xls"}],"financialInterests":"No competing interests reported.","formattedTitle":"Novel drug resistance- and macrophage polarization-related molecular subtyping and prognostic signature for pancreatic adenocarcinoma","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003ePancreatic adenocarcinoma (PAAD) is recognized as a highly aggressive form of solid malignancy, characterized by a notably poor clinical outcome. Current treatment modalities for PAAD are limited and primarily include surgical resection, chemotherapy, radiotherapy, and immunotherapy [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. However, the efficacy of these approaches is often compromised due to the emergence of drug resistance. Therefore, identifying novel biomarkers is critical for stratifying PAAD patients and assessing the prognosis of distinct patient subgroups [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe PAAD tumor microenvironment (TME) exhibits a dense extracellular matrix enriched with activated fibroblasts and immune cells [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. A dense stroma impairs vascularization and vessel functionality, thereby hindering the effective delivery of therapeutic agents. Tumor-associated macrophages (TAMs), as pivotal cellular constituents within the tumor microenvironment (TME), are key players in the growth, recurrence, and spread of PAAD tumors and in modulating immune responses and reactions to chemotherapy and immunotherapy [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Elevated infiltration of TAMs are in connection with poor therapeutic outcomes in PAAD, and reducing the number of TAMs has been shown to suppress pancreatic tumor development [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Nevertheless, the intrinsic factors governing macrophage development and function in PAAD remain largely unknown, which significantly limits the therapeutic potential of TAMs as targets and tools for PAAD treatment. Prior studies have emphasized the fundamental importance of DMRGs in modulating the tumor microenvironment and contributing to therapeutic resistance [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Despite these advances, a significant gap remains in the identification of distinct pancreatic cancer subtypes and development of robust prognostic models that integrate these genetic factors. Therefore, this study discovered novel biomarkers and therapeutic targets to enhance prognostic accuracy and ultimately improve clinical outcomes in the management of PAAD.\u003c/p\u003e \u003cp\u003eTo this end, we employed a comprehensive bioinformatics approach incorporating techniques such as batch effect removal, consistent clustering analysis, differential expression, and Cox regression modeling. The advantage of this multifaceted approach is its ability to integrate information from multiple datasets, thereby enhancing the reliability and reproducibility of the results. The primary objective of this study was to identify the distinct subtypes of pancreatic cancer and investigate the expression characteristics of DMRGs. Through the development of a comprehensive prognostic scoring system, we aimed to elucidate the pivotal roles of these molecular markers in PAAD. This approach facilitates the discovery of novel diagnostic biomarkers and potential therapeutic interventions, ultimately improving patient prognosis and treatment efficacy.\u003c/p\u003e"},{"header":"2 Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Data download\u003c/h2\u003e \u003cp\u003eThe PAAD dataset (TCGA-PAAD), was obtained from TCGA database and utilized for analytical evaluation. Samples lacking clinical information were excluded and then we obtained sequencing data in count format for 176 PAAD samples with clinical information. Meanwhile, the data was accessed through the UCSC Xena database to acquire corresponding clinical data (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBaseline Table with PAAD Patients Characteristics\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"2\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003echaracteristics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eoverall\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFEMALE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e79 (44.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMALE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e97 (55.1%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStage_M, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eM0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e79 (44.9%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eM1\u0026amp;MX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e97 (55.1%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStage_N, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e49 (27.8%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN1\u0026amp;NX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e127 (72.2%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStage_T, n (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT1\u0026amp;T2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e30 (17%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT3\u0026amp;T4\u0026amp;TX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e146 (83%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"2\"\u003ePAAD, Pancreatic adenocarcinoma\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe PAAD datasets GSE28735 and GSE71729 were obtained from the Gene Expression Omnibus (GEO) database using the GEOquery package (version 2.70.0) [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Dataset GSE28735 uses the GPL6244 chip platform, while dataset GSE71729 utilizes GPL20769. More details are available in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. The dataset GSE28735 contains 45 PAAD and 45 control samples. The dataset GSE71729 includes 145 PAAD and 46 control samples.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eGEO Microarray Chip Information\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGSE28735\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGSE71729\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlatform\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGPL6244\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGPL20769\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSpecies\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHomo sapiens\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHomo sapiens\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTissue\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePancreas\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePancreas\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSamples in PAAD group\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e145\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSamples in Control group\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e46\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePMID: 22363658\u003c/p\u003e \u003cp\u003ePMID: 23918603\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePMID: 26343385\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"3\"\u003eGEO, Gene Expression Omnibus;PAAD, Pancreatic adenocarcinoma\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThrough comprehensive mining of the GeneCards [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e] and DRESIS [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e] databases, we compiled a catalog of genes implicated in both drug resistance mechanisms and macrophage polarization processes (hereafter referred to as DMRGs). We first downloaded the drug resistance-related gene set from the DRESIS database, retaining only the drug resistance-related genes (DRGs) with molecule type as 'Protein' and Molecule species as 'Homo sapiens.\u0026rsquo; A systematic literature search was conducted on PubMed using \"drug resistance\" as the primary search term to identify relevant published studies [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. The set of DRGs was consolidated and de-duplicated, yielding 959 DRGs. Similarly, we identified 586 macrophage polarization-related genes (MRGs) [\u003cspan additionalcitationids=\"CR13 CR14\" citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. After determining the intersection of genes related to the two phenotypes of drug resistance and macrophage polarization, 106 DMRGs were obtained (Supplementary material 4 - Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe batch effect correction was implemented on GSE28735 and GSE71729 datasets through the sva package (v3.50.0), resulting in a consolidated GEO dataset containing 190 PAAD specimens. To evaluate the normalization efficacy, expression profile distributions were visualized and compared using boxplot representations prior to and following the standardization procedure.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Construction of PAAD subtypes\u003c/h2\u003e \u003cp\u003eTo identify potential prognostic biomarkers in PAAD from the TCGA dataset, univariate Cox proportional hazards regression was conducted incorporating clinical parameters. Following this initial screening, we employed consensus clustering methodology to stratify PAAD cases into molecularly distinct subgroups characterized by differential expression patterns of the identified prognostic genes.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 DEGs analysis\u003c/h2\u003e \u003cp\u003eBased on the grouping of PAAD samples from TCGA-PAAD, gene expression analysis was conducted across various disease subtypes. Differential gene expression analysis identified statistically significant genes (DEGs) using stringent criteria: an absolute log2 fold change greater than 1 (|log2FC| \u0026gt; 1) combined with an adjusted p-value below 0.05.\u003c/p\u003e \u003cp\u003eAn intersection analysis between DEGs and a set of DMRGs was performed and the overlapping genes yielded the DMRDEGs. The 20 most significant differentially DMRDEGs were visualized through pheatmap package (v1.0.12) in R. Furthermore, their genomic distribution across chromosomes was mapped and presented using the RCircos package (v1.2.2).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Prognosis analysis of pancreatic cancer\u003c/h2\u003e \u003cp\u003eKaplan-Meier (KM) survival analysis was performed to assess survival outcome disparities between different molecular subtypes of PAAD. Additionally, the ggplot2 package was utilized to generate visual representations of clinical feature distributions across patient subgroups, enabling systematic evaluation of potential prognostic determinants.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Analysis of somatic mutations (SMs) and copy number variations (CNVs)\u003c/h2\u003e \u003cp\u003eFor SM data processing, we employed VarScan for initial quality control and variant calling. Subsequently, we performed comprehensive characterization of CNV patterns, including their genomic distribution and prevalence, through visualization analyses implemented in ggplot2 (v3.4.4).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.6 Immunotherapy analysis\u003c/h2\u003e \u003cp\u003eWe sourced 25 immune checkpoint genes (ICGs) from the PubMed literature [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e], and their names are detailed in Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e (Supplementary material 5). The expression differences of these ICGs in PAAD samples were analyzed, and grouped comparison charts were generated.\u003c/p\u003e \u003cp\u003eThe immune infiltration scores for PAAD samples were calculated using transcriptomic data obtained from the TCGA-PAAD cohort through the TIDE online platform [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.7 Gene set enrichment analysis (GSEA)\u003c/h2\u003e \u003cp\u003eGSEA was performed on the comprehensive RNA sequencing dataset obtained from TCGA-PAAD cohort, utilizing version 4.10.0 of the clusterProfiler R package. The analytical parameters were carefully selected to optimize both statistical power and biological significance: a fixed random seed (2020) was implemented to guarantee reproducibility, 1000 permutations were executed to establish robust statistical inference, and gene sets were restricted to those comprising between 10 and 500 genes to maintain biological interpretability. Pathway annotations were obtained from the Molecular Signatures Database (MSigDB) [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. Statistically significant pathways were determined using strict criteria: Benjamini-Hochberg corrected p-values less than 0.05 and false discovery rates (q-values) below 0.25, which provided robust control for multiple comparisons.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e2.8 Development of a predictive risk stratification model for PAAD\u003c/h2\u003e \u003cp\u003eInitially, we recognized prognostic factors demonstrating statistical significance (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05). These significant variables were subjected into a model for further evaluation.\u003c/p\u003e \u003cp\u003eSubsequently, to construct the prognostic model, we employed Least Absolute Shrinkage and Selection Operator (LASSO) regression analysis with the family parameter specified as \"cox\". This analytical approach utilized DMRDEGs identified in the previous analysis. Through 10-fold cross-validation, we determined the optimal set of predictive genes for the risk assessment model. The results of this LASSO regression analysis were visually represented through two distinct graphical outputs: a risk score prognostic plot and a coefficient profile trajectory plot.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e2.9 Evaluation of PAAD risk prediction model\u003c/h2\u003e \u003cp\u003eThe predictive model's ability to distinguish temporal patterns was evaluated by conducting time-dependent receiver operating characteristic (ROC) curve analysis, implemented using the survivalROC package (version 1.0.3.1), which enabled the computation of area under the curve (AUC) values for 1-, 2-, and 3-year survival predictions. Survival analysis was conducted using the survival package (v3.5-7) to generate Kaplan-Meier curves. For clinical application, we developed a prognostic nomogram through the rms package (v6.7-1), incorporating significant covariates identified.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e2.10 Immune infiltration analysis\u003c/h2\u003e \u003cp\u003eThe differential expression patterns of immune cells across subtypes (Cluster 1 vs. Cluster 2) were visualized using ggplot2 (v3.4.4). For comprehensive representation of correlation analyses, pheatmap (v1.0.12) facilitated the generation of heatmaps depicting both intercellular relationships among immune cells and associations between model genes and immune cell populations. This dual visualization approach enabled systematic examination of immune microenvironment characteristics.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e2.11 Validation of differential expression and analysis of ROC Curve\u003c/h2\u003e \u003cp\u003eWe determined risk scores for each PAAD sample across the integrated GEO datasets based on established risk coefficients. A comparative comparison chart was generated to visualize the differential expression patterns of the model genes across the two risk groups. We constructed a comparative chart to illustrate the distinct transcriptional profiles of signature genes between high-risk and low-risk patient cohorts.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e2.12 Construction of protein-protein interaction networks and identification of hub genes\u003c/h2\u003e \u003cp\u003eEmploying the STRING database [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e], we established a PPI network comprising model genes and subsequently identified candidate genes from this network for additional investigation. Furthermore, the GeneMANIA web-based tool [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e] was utilized to detect genes exhibiting functional associations with hub genes implicated in drug resistance and macrophage polarization.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e2.13 Construction of regulatory networks\u003c/h2\u003e \u003cp\u003eTo investigate transcription factor-mediated (TF-mediated) gene regulation, we systematically examined the regulatory interactions between transcription factors and their target genes using data obtained from the ChIPBase repository [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. we obtained miRNAs associated with model genes and predicted the target RNA-Binding Protein (RBP) for model genes Using StarBase v3.0 database [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e2.14 ESTIMATE analysis\u003c/h2\u003e \u003cp\u003eWe calculated ESTIMATE immune and stromal scores with R package IOBR (version 0.99.9). And we generated comparative plots for three distinct immune infiltration analyses.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e2.15 Statistical analysis\u003c/h2\u003e \u003cp\u003eFor the comparison of continuous variables across two groups, parametric data were assessed via independent two-sample t-tests, whereas nonparametric distributions were analyzed using Mann-Whitney U tests. Bivariate associations were quantified through Spearman's rank correlation coefficients. The complete statistical workflow was implemented using R statistical computing environment (version 4.3.0).\u003c/p\u003e \u003c/div\u003e"},{"header":"3 Results","content":"\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Two subtypes were identified in PAAD based on DMRGs\u003c/h2\u003e \u003cp\u003eThe methodological pipeline for the systematic investigation of DMRGs is depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA. Initially, we eliminated batch effects between the GSE28735 and GSE71729 PAAD datasets with \u0026ldquo;sva\u0026rdquo; package, resulting in an integrated GEO dataset. The comparative analysis of expression profiles, as visualized through distribution box plots, revealed distinct patterns between pre- and post-batch correction datasets (Supplementary material 1 - Fig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003eA and S1B). To further evaluate the intrinsic structure of the integrated dataset, principal component analysis (PCA) was employed to visualize the distribution patterns in reduced dimensions (Supplementary material 1 - Fig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003eC and S1D). The graphical representations clearly demonstrate the effectiveness of the normalization procedure in reducing technical variability while preserving biological signals. Among the 106 DMRGs, 36 were identified as significantly associated with prognosis based on univariate Cox regression analysis, and their hazard ratios were graphically represented in a forest plot (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB).\u003c/p\u003e \u003cp\u003eAccording to the expression profiles of 36 DMRGs, a consensus clustering analysis was performed to identify distinct molecular subtypes. The analytical results delineated two distinct molecular subgroups of PAAD: Subtype A (Cluster 1), which encompassed 162 tumor specimens, and Subtype B (Cluster 2), containing 14 cases (Figs.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC, \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD, and Supplementary material 2 - Fig. \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). This classification was consistently supported by multiple analytical approaches. The three-dimensional t-SNE clustering plot demonstrated distinct clustering patterns that clearly differentiated between the two subtypes (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eE).\u003c/p\u003e \u003cp\u003eTo investigate potential variations in gene expression profiles and their associations with biological properties, disease pathogenesis, and clinical phenotypes, we performed comparative transcriptomic analysis. Applying stringent statistical criteria (|log2 fold change| \u0026gt; 1 with adjusted p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05), our differential expression analysis of TCGA-PAAD samples identified 11,176 significantly dysregulated genes. Among these, 4,521 transcripts exhibited increased expression levels while 6,655 showed decreased expression patterns, as illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eF.\u003c/p\u003e \u003cp\u003eThrough comprehensive analysis, we identified 34 DMRDEGs from the intersection of 11,142 DEGs and 36 DMRGs, as illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eG. Complete annotation data for these candidate genes are provided in Table \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e (Supplementary material 6). Statistical analysis revealed significant differential expression patterns among the top 20 DMRDEGs when ranked according to the magnitude of their log2 fold-change values in the TCGA-PAAD dataset (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eH). To conclude our analysis, we employed the RCircos package in R to conduct chromosomal localization studies, which enabled precise mapping of gene loci across the genome. This computational approach yielded a detailed visualization of genomic distribution patterns, as illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eI. And the results indicated that the 34 DMRDEGs mainly located on chromosome 1, specifically, \u003cem\u003ePIK3CD\u003c/em\u003e, \u003cem\u003eSLC2A1\u003c/em\u003e, \u003cem\u003eS100A8\u003c/em\u003e, \u003cem\u003eMUC1\u003c/em\u003e, and \u003cem\u003ePTGS2\u003c/em\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Characteristics of the two PAAD subtypes\u003c/h2\u003e \u003cp\u003eOur GSEA analysis demonstrated substantial enrichment across multiple biological processes and signaling cascades (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). Notably, we observed prominent activation of the Fc epsilon receptor (Fceri) signaling pathway (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB), along with FCGR3A-dependent interleukin-10 (\u003cem\u003eIL-10\u003c/em\u003e) production (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC). The study further identified Fceri-triggered calcium ion flux (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD) and the participation of the \u003cem\u003eLAT2/NTAL\u003c/em\u003e mechanism in calcium signaling (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eE).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eResults of GSEA for PAAD\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"9\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003eID\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSet Size\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eEnrichment Score\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNES\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c7\" namest=\"c6\"\u003e \u003cp\u003epvalue\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003ep.adjust\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eqvalue\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eREACTOME_ROLE_OF_LAT2_NTAL_LAB_ON_CALCIUM_MOBILIZATION\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e71\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.879869\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e3.280909\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.00E-10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e5.24E-09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e3.75E-09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eREACTOME_FCERI_MEDIATED_CA_2_MOBILIZATION\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e86\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.836874\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e3.244674\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.00E-10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e5.24E-09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e3.75E-09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eREACTOME_FCGR3A_MEDIATED_IL10_SYNTHESIS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e94\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.784836\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e3.106986\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.00E-10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e5.24E-09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e3.75E-09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eREACTOME_FC_EPSILON_RECEPTOR_FCERI_SIGNALING\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e186\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.657411\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.884516\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.00E-10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e5.24E-09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e3.75E-09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eREACTOME_ASSEMBLY_OF_COLLAGEN_FIBRILS_AND_OTHER_MULTIMERIC_STRUCTURES\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.725634\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.672979\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.00E-10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e5.24E-09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e3.75E-09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWP_OVERVIEW_OF_PROINFLAMMATORY_AND_PROFIBROTIC_MEDIATORS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e115\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.654766\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.651993\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.00E-10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e5.24E-09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e3.75E-09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eREACTOME_INTERLEUKIN_10_SIGNALING\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.757527\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.618502\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.00E-10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e5.24E-09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e3.75E-09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eREACTOME_INTERLEUKIN_4_AND_INTERLEUKIN_13_SIGNALING\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e111\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.560062\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.264255\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e8.85E-10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e3.91E-08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e2.79E-08\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eREACTOME_MET_PROMOTES_CELL_MOTILITY\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.657049\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.234189\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.10E-06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e2.34E-05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e1.67E-05\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eREACTOME_NEUTROPHIL_DEGRANULATION\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e475\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.463173\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.230004\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.00E-10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e5.24E-09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e3.75E-09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eREACTOME_SIGNALING_BY_INTERLEUKINS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e470\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.463849\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.229662\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.00E-10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e5.24E-09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e3.75E-09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eREACTOME_ANTI_INFLAMMATORY_RESPONSE_FAVOURING_LEISHMANIA_PARASITE_INFECTION\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e211\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.497813\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.211216\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.00E-10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e5.24E-09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e3.75E-09\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eREACTOME_TNF_RECEPTOR_SUPERFAMILY_TNFSF_MEMBERS_MEDIATING_NON_CANONICAL_NF_KB_PATHWAY\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.79427\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.179953\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e5.80E-06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.000104\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e7.40E-05\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eREACTOME_DECTIN_2_FAMILY\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.703239\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.15809\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.05E-05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.000171\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.000122\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eREACTOME_CROSSLINKING_OF_COLLAGEN_FIBRILS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e18\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.77486\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.151253\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e2.82E-05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.000406\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.00029\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePID_AMB2_NEUTROPHILS_PATHWAY\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.623443\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.119918\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e9.98E-06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.000167\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.000119\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePID_PI3KCI_PATHWAY\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.59446\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.089981\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.73E-05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.000271\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.000194\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBIOCARTA_IL1R_PATHWAY\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.657375\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.087755\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e3.19E-05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.00045\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.000322\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eREACTOME_ANCHORING_FIBRIL_FORMATION\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.776326\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.076566\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e9.39E-05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.001098\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.000785\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eREACTOME_COSTIMULATION_BY_THE_CD28_FAMILY\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003e72\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.551565\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e2.065634\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.16E-06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e2.42E-05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e1.73E-05\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"9\"\u003eGSEA, Gene Set Enrichment Analysis\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eKM survival analysis revealed that subtypes A (cluster 1) has a highly poorer OS than subtypes B (cluster 2) (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eF). Furthermore, we analyzed the proportions of patients in subtype 1 and subtype 2 across Age, Gender, Stage_M, Stage_N, and Stage_T categories. The figure indicates that more PAAD patients in subtypes A (cluster 1) are in the worse T3, T4 and TX stages than B (cluster 2) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eG).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Immune environments differ between the two PAAD subtypes\u003c/h2\u003e \u003cp\u003eEmploying the CIBERSORT computational method, we estimated the proportions of 22 distinct immune cell populations within the TCGA-PAAD cohort. Five distinct immune cell populations\u0026mdash;regulatory T cells (Tregs), activated natural killer (NK) cells, monocytes, M0 macrophages, and M1 macrophages\u0026mdash;exhibited statistically significant variations across the identified subtypes (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05; Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA).\u003c/p\u003e \u003cp\u003eSubsequently, a heatmap was employed to illustrate the associations among the infiltration levels of these five immune cell types. The results showed that most immune cells exhibited a strong correlation in these two subtypes, with the strongest significant negative correlation between monocytes and M0 macrophages in subtype A (r = -0.418, p\u0026thinsp;\u0026lt;\u0026thinsp;0.05; Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB), and the robust inverse association between monocytes and M0 macrophages in subtype B (r = -0.55, p\u0026thinsp;\u0026lt;\u0026thinsp;0.05; Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC).\u003c/p\u003e \u003cp\u003eICGs were identified based on previously published studies and publicly accessible databases. Through cross-referencing these genes with the TCGA-PAAD database, we obtained an expression profile matrix containing 25 ICGs and their respective transcriptional quantities. Subsequently, we analyzed the differential expression patterns of ICGs between subtype A (Cluster 1) and subtype B (Cluster 2), as illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD. Significant differences were observed between subtypes A and B for most ICGs (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), including \u003cem\u003eFCRL4, G0S2, CD38, CLEC7A, CCL13, CCR2, CXCL10, CD1E, LILRB2, CD47, CD70, CXCL1, HLA-DOB, CXCL11, LILRB1, CCL18, CCR7, LGALS9\u003c/em\u003e and \u003cem\u003eBIRC3\u003c/em\u003e. Our investigation focused on assessing the responsiveness of PAAD patients to immunotherapeutic interventions, leveraging this computational approach (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eE). The results indicated that subtype B may exhibit better responsiveness to immunotherapy than subtype A.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Analysis of CNVs and SMs in DMRDEGs\u003c/h2\u003e \u003cp\u003eTo investigate SM profiles of the 34 DMRDEGs in PAAD samples from the TCGA-PAAD cohort, we conducted comprehensive mutational profiling using the \"maftools\" R package. Our analysis revealed seven principal SM categories affecting these DMRDEGs, with missense mutations representing the most prevalent alteration type (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA, left panel). Additionally, the primary mutation type observed among the 34 DMRDEGs within the PAAD group was single nucleotide polymorphisms (SNPs) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA, middle panel). The predominant single-nucleotide variations (SNVs) identified in PAAD specimens were characterized by cytosine-to-thymine transitions, as demonstrated in the right panel of Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA right panel. Furthermore, we analyzed the SM profiles of the 34 DMRDEGs and ranked them based on mutation frequency, followed by visual representation (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). \u003cem\u003eANXA1\u003c/em\u003e and \u003cem\u003ePIK3CG\u003c/em\u003e exhibited the highest mutation rates and their mutation frequency was 2%.\u003c/p\u003e \u003cp\u003eTo analyze the CNVs in the 34 DMRDEGs, we downloaded and synthesized CNV data from the PAAD group. Using GISTIC 2.0, CNVs of 34 DMRDEGs in the PAAD group were identified, and the CNV profiles of these genes were visualized (Figs.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC and D).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec23\" class=\"Section2\"\u003e \u003ch2\u003e3.5 Development of a predictive risk stratification framework for PAAD\u003c/h2\u003e \u003cp\u003eA prognostic risk model for PAAD was constructed with 34 DMRDEGs. All significant variables (p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05) were visualized using a forest plot (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA). Through comprehensive analysis, 19 DMRDEGs exhibiting significant prognostic relevance (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05) were identified. To further analysis, a LASSO regression model was developed (Figs.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB and \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC). The final model incorporated five key genes: \u003cem\u003eIL18\u003c/em\u003e, \u003cem\u003eEREG\u003c/em\u003e, \u003cem\u003eLDHA\u003c/em\u003e, \u003cem\u003eSOCS2\u003c/em\u003e, and \u003cem\u003eSPP1\u003c/em\u003e. The risk score was derived based on the following equation:\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:RiskScore\\:=\\:IL18\\:\\ast\\:\\:\\left(0.112\\right)\\:+\\:EREG\\:\\ast\\:\\:\\left(0.05\\right)\\:+\\:LDHA\\:\\ast\\:\\:\\left(0.256\\right)\\:+\\:SOCS2\\:\\ast\\:\\:(-0.231)\\:+\\:SPP1\\:\\ast\\:\\:\\left(0.028\\right)\u0026quot;$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eTo investigate the functional associations between the five key genes identified in our model, we conducted a PPI analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD). Building upon these findings, we developed a comprehensive interaction network that integrates these hub genes involved in both drug resistance mechanisms and macrophage polarization processes, along with their functionally related counterparts (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eE). The resulting network architecture included the five core model genes and an additional set of 20 proteins sharing similar biological functions.\u003c/p\u003e \u003cp\u003eTo elucidate the molecular mechanisms governing transcriptional regulation, we constructed a comprehensive mRNA-TF interaction network. This network was subsequently visualized and analyzed using the Cytoscape platform, enabling systematic examination of potential regulatory relationships (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eF). This regulatory network incorporated four key model genes along with 36 TF, with comprehensive data available in Supplementary Table \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e (Supplementary material 7). Subsequently, we predicted microRNAs (miRNAs) that potentially interact with these model genes, leading to the construction of an mRNA-miRNA regulatory network. This secondary network encompassed three model genes and 43 miRNAs (Supplementary material 3 - Fig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e), with complete annotation details provided in Supplementary Table \u003cspan refid=\"MOESM5\" class=\"InternalRef\"\u003eS5\u003c/span\u003e (Supplementary material 8). Finally, we constructed an mRNA\u0026ndash;RBP regulatory network to visualize RBPs associated with the model genes and the network comprised four model genes and 42 RBPs (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eG and Supplementary material 9 - Table \u003cspan refid=\"MOESM6\" class=\"InternalRef\"\u003eS6\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFurthermore, Statistical evaluation of the association between the predictive gene markers and immune cell infiltration patterns demonstrated that the majority of immune cell types showed significant associations with the five model genes in both subtype A and subtype B. And \u003cem\u003eSOCS2\u003c/em\u003e had the strongest negative correlation with M0 macrophages in subtype A (r = -0.458, p\u0026thinsp;\u0026lt;\u0026thinsp;0.05; Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eH), while \u003cem\u003eSOCS2\u003c/em\u003e had the strongest significant positive correlation with activated NK cells in subtype B (r\u0026thinsp;=\u0026thinsp;0.626, p\u0026thinsp;\u0026lt;\u0026thinsp;0.05; Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eI).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003e3.6 Clinical evaluation of the prognostic risk model\u003c/h2\u003e \u003cp\u003eSubsequently, we constructed ROC curves utilizing the TCGA-PAAD cohort to evaluate the predictive performance of our prognostic signature (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA). This analytical approach enabled us to quantitatively assess the model's discrimination capacity at various time points during follow-up. The results showed that the model demonstrated moderate to high accuracy at 1 and 3 years (0.7\u0026thinsp;\u0026lt;\u0026thinsp;AUC\u0026thinsp;\u0026lt;\u0026thinsp;0.9), with slightly lower accuracy at 2 years (0.5\u0026thinsp;\u0026lt;\u0026thinsp;AUC\u0026thinsp;\u0026lt;\u0026thinsp;0.7). KM survival analysis demonstrated a statistically significant divergence in overall survival between high- and low-risk patient cohorts (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01; Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB). The nomogram constructed to represent the association between risk scores and five key clinical variables, indicated that the riskscore exhibited substantially greater predictive value within the prognostic model compared to the other variables (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eC).\u003c/p\u003e \u003cp\u003eSubsequently, we conducted a univariate Cox proportional hazards regression analysis utilizing the median RiskScore as the stratification threshold, incorporating OS data and relevant clinical parameters extracted from the database. The results from both univariate (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eD) and multivariate (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eE) regression analyses were illustrated using forest plots, and detailed statistical parameters are provided in Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. Our analysis revealed that the RiskScore demonstrated remarkable statistical significance (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) in predicting patient outcomes.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eResults of Cox Analysis\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eCharacteristics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eTotal(N)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003eUnivariate analysis\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003eMultivariate analysis\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHR (95% CI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eP value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHR(95% CI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eP value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e176\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.025 (1.003\u0026ndash;1.047)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.025 (1.004\u0026ndash;1.046)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.021\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e176\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFEMALE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMALE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e97\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.788 (0.522\u0026ndash;1.190)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.257\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStage_M\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e176\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eM0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eM1\u0026amp;MX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e97\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.851 (0.562\u0026ndash;1.287)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.444\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStage_N\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e176\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eN1\u0026amp;NX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e127\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.991 (1.188\u0026ndash;3.337)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.009\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.727 (1.019\u0026ndash;2.929)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.043\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStage_T\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e176\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT1\u0026amp;T2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eReference\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eT3\u0026amp;T4\u0026amp;TX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e146\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.781 (0.946\u0026ndash;3.356)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.074\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.176 (0.616\u0026ndash;2.245)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.624\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRisk.Score\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e176\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.158 (2.685\u0026ndash;9.910)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5.182 (2.668\u0026ndash;10.062)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eHR, Hazard ratio. An HR\u0026thinsp;\u0026gt;\u0026thinsp;1 indicates that the variable is a risk factor, while an HR\u0026thinsp;\u0026lt;\u0026thinsp;1 suggests that it is a protective factor. Variables with a univariate p value\u0026thinsp;\u0026lt;\u0026thinsp;0.1 were included in the analysis\u003c/p\u003e \u003cp\u003eThe predictive accuracy of our risk stratification model was systematically evaluated through calibration analyses conducted at annual intervals (1-, 2-, and 3-year follow-up periods), with the resulting calibration plots presented in Figs.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eF through \u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eH. The results indicated that the model exhibited optimal predictive accuracy for clinical outcomes at the 1-year interval. Next, we used DCA at the same time intervals to assess the clinical utility (Figs.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eI\u0026ndash;K). Our evaluation revealed that the LASSO regression model exhibited progressively enhanced predictive accuracy for longer-term outcomes, with the highest performance observed for 3-year predictions, followed by 2-year and then 1-year forecasts.\u003c/p\u003e \u003cp\u003eTo investigate the expression of the model genes between groups, we used a grouped comparison chart to illustrate the expression analysis results (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eL). Statistical analysis revealed markedly distinct expression patterns of the five signature genes between high- and low-risk cohorts, with statistical significance (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e \u003cp\u003eSubsequently, ROC curve analysis was performed to evaluate the predictive performance of selected model genes The results revealed that \u003cem\u003eLDHA\u003c/em\u003e expression displayed excellent discriminative capacity (AUC\u0026thinsp;\u0026gt;\u0026thinsp;0.9, Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eN), while \u003cem\u003eIL18, EREG\u003c/em\u003e and \u003cem\u003eSOCS2\u003c/em\u003e expressions exhibited intermediate predictive accuracy (0.7\u0026thinsp;\u0026le;\u0026thinsp;AUC\u0026thinsp;\u0026lt;\u0026thinsp;0.9, Figs.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eM and \u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eP). In contrast, SPP1 showed relatively limited discriminatory power (0.5\u0026thinsp;\u0026lt;\u0026thinsp;AUC\u0026thinsp;\u0026lt;\u0026thinsp;0.7, Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eQ) in distinguishing between high- and low-risk patient subgroups.\u003c/p\u003e \u003cp\u003eFurthermore, we quantified the stromal, immune, and ESTIMATE scores based on the transcriptomic profiles from the TCGA-PAAD cohort. Comparative analysis between high- and low-risk patient subgroups revealed statistically significant disparities in stromal compartment scores (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01), as illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eO. These findings demonstrate distinct tumor microenvironment characteristics between the prognostic groups. The immunological and ESTIMATE scoring systems exhibited highly significant differences (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) when comparing the two cohorts, with notably elevated values observed in patients classified as low-risk relative to their high-risk counterparts.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"4 Discussion","content":"\u003cp\u003eSubtype-specific molecular traits can influence treatment responses and clinical outcomes, suggesting that a deeper understanding of these variations could improve therapeutic efficacy [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. PAAD represents a highly heterogeneous malignancy characterized by the presence of distinct molecular subtypes. These subtypes exhibit notable differences in prognosis and treatment response [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. Statistical analyses reveal that pancreatic adenocarcinoma (PAAD) patients exhibit a dismal prognosis, with fewer than 15% surviving beyond five years post-diagnosis, largely due to the development of drug resistance. Although initial treatments may be effective, resistance often develops rapidly, significantly diminishing therapeutic success [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. Within the TME, macrophages serve as pivotal mediators that significantly influence both neoplastic advancement and therapeutic resistance. These cell populations demonstrate significant adaptability, enabling their differentiation into specialized functional subtypes, notably the conventionally activated M1 and alternatively stimulated M2 phenotypes. M1-polarized macrophages are primarily associated with pro-inflammatory responses and exhibit anti-tumor properties, while M2-polarized macrophages tend to promote immunosuppressive functions and facilitate tumor progression [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. The involvement of macrophage polarization in mediating drug resistance in PAAD has received increasing attention [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. Recent studies have demonstrated that modulating macrophage polarization can affect cancer cell sensitivity to various therapies [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. Therefore, macrophage polarization may be linked to drug resistance in PAAD.\u003c/p\u003e \u003cp\u003eWe initially identified 36 DMRGs from public databases and based on these genes, we distinguished two PAAD subtypes: A (cluster 1) and B (cluster 2). These subtypes exhibit significant clinical differences, with subtype A showing a considerably better prognosis than subtype B (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). GSEA revealed that PAAD samples were enriched in immune responses and cell signaling pathways, emphasizing the critical role of the TME in cancer progression (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Specifically, pathways such as Fceri signaling and FCGR3A-mediated IL-10 production highlight the complex interactions between tumor cells and immune components. Within the TME, immune complexes formed by tumor antigens and IgG antibodies bind to FCGR3A on TAMs, activating the Syk/PI3K/Akt or nuclear factor kappa-beta pathways and promoting IL-10 gene transcription [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. Secreted IL-10 can upregulate M2 markers, thus promoting TAM polarization towards an M2 phenotype associated with tumor progression [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. Animal models have demonstrated that knocking out FCGR3A or blocking IL-10 reduces M2-type TAM infiltration, inhibiting tumor growth and metastasis [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. Therefore, targeting this pathway holds promise for reversing the tumor-promoting effects of TAMs and for developing new strategies for tumor immunotherapy. However, further investigation is needed to elucidate the mechanistic differences across tumor contexts and optimize precision intervention strategies.\u003c/p\u003e \u003cp\u003eThe identification of individuals with a higher likelihood of positive response to immunotherapeutic interventions is crucial for improving clinical outcomes [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. Notably, a substantial difference in TIDE immunotherapy scores existed between subtypes A and B, indicating that subtype B may be more responsive to immune checkpoint inhibitors than subtype A (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Immunological profiling using CIBERSORT demonstrated significant intercellular associations within PAAD subtype A, where a marked inverse relationship was observed between monocyte populations and M0 macrophage subsets (r = -0.418, p\u0026thinsp;=\u0026thinsp;0.032). This analysis highlighted the dynamic interplay between distinct immune cell lineages in PAAD microenvironment. Notably, subtype B exhibited pronounced inverse relationships among immune cell populations, particularly between monocytes and M0 macrophages which displayed the most substantial negative correlation (r = -0.55, p-value\u0026thinsp;=\u0026thinsp;0.047). These immunological profiling results provide crucial insights for designing precision immunotherapeutic interventions customized to patients with distinct characteristics, ultimately leading to optimized personalized treatment plans and a reduced patient burden.\u003c/p\u003e \u003cp\u003eTo further validate the prognostic significance of DEGs across the PAAD subtypes, we performed LASSO regression analysis using a panel of 19 DMRDEGs and developed a corresponding risk model (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e). This model could predict PAAD survival rates, paving the way for tailored immunotherapy for patients with PAAD who are susceptible to drug resistance. The model comprised five genes: IL18, EREG, LDHA, SOCS2, and SPP1. CIBERSORT analysis revealed that in subtype A, SOCS2 expression was negatively related to the upregulation of M0 macrophages (r = -0.458, p\u0026thinsp;\u0026lt;\u0026thinsp;0.05), suggesting that SOCS2 negatively regulates M0 macrophage aggregation. SOCS2, a member of the SOCS family, regulates signaling pathways such as JAK-STAT, and is involved in tumorigenesis, development, and drug resistance [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. Mechanistically, SOCS2 hinders tumor cell survival and invasion by inhibiting the JAK2/STAT5 pathway [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. In the immune microenvironment, SOCS2 decreases M2-type TAM polarization and immunosuppressive factor secretion by inhibiting IL-6/JAK2/STAT3 signaling [\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. In malignancies exhibiting reduced SOCS2 levels, particularly in non-small cell lung carcinoma cases, experimental restoration of SOCS2 expression has been demonstrated to suppress metastatic potential and reverse chemotherapeutic resistance [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. Similarly, in PAAD, SOCS2 may contribute to the resistance to chemotherapy and targeted therapy via related mechanisms. Restoring SOCS2 function or targeting its downstream pathways may emerge as novel strategies for overcoming drug resistance, pending further translational research. SOCS2 mRNA and protein expression is significantly downregulated in PAAD cells compared to normal pancreatic cells [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. Nevertheless, the functional significance of SOCS2 in mediating chemoresistance mechanisms within PAAD has not been fully elucidated. In future, we aim to investigate the dynamic regulatory network of SOCS2 in PAAD drug resistance using single-cell sequencing and organoid models.\u003c/p\u003e \u003cp\u003eThis investigation presents several limitations that warrant consideration. The modest cohort size may constrain the broader applicability of the results. Furthermore, the lack of experimental validation in laboratory settings limits the confirmation of proposed biomarkers and their mechanistic involvement in pancreatic adenocarcinoma. The integration of diverse datasets also raises the possibility of batch-related variations, which could influence the consistency and clarity of the observed outcomes.\u003c/p\u003e"},{"header":"5 Conclusions","content":"\u003cp\u003eIn conclusion, our analysis categorized PAAD into two distinct subgroups, designated as Subtype A (1) and Subtype B (2), based on DMRGs. And we also successfully identified 34 DMRDEGs that are crucial for understanding the TME in patients. A prognostic risk model developed from these genes exhibits considerable promise for forecasting patient survival outcomes. Future research could pay attention on clarifying the regulatory mechanisms of macrophage polarization and validating these findings in larger clinically relevant cohorts. Finally, this study lays the groundwork for the development of novel therapeutic targets and strategies to improve PAAD treatment outcomes.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003ePAAD, pancreatic adenocarcinoma\u003c/p\u003e\n\u003cp\u003eTCGA, The Cancer Genome Atlas\u003c/p\u003e\n\u003cp\u003eGEO, Gene Expression Omnibus\u003c/p\u003e\n\u003cp\u003eDMRGs, drug resistant and macrophage polarization-related genes\u003c/p\u003e\n\u003cp\u003eDEGs, differentially expressed genes\u003c/p\u003e\n\u003cp\u003eDMRDEGs, drug resistance- and macrophage polarization-related differentially expressed genes\u003c/p\u003e\n\u003cp\u003eTME, tumor microenvironment\u003c/p\u003e\n\u003cp\u003eTAM, tumor-associated macrophage\u003c/p\u003e\n\u003cp\u003eFPKM, fragments per kilobase per million\u003c/p\u003e\n\u003cp\u003eDRGs, drug resistance-related genes\u003c/p\u003e\n\u003cp\u003eMRGs, macrophage polarization-related genes\u003c/p\u003e\n\u003cp\u003eOS, overall survival\u003c/p\u003e\n\u003cp\u003eKM, Kaplan\u0026ndash;Meier\u003c/p\u003e\n\u003cp\u003eSM, somatic mutations\u003c/p\u003e\n\u003cp\u003eICGs, immune checkpoint genes\u003c/p\u003e\n\u003cp\u003eCNV, copy number variations\u003c/p\u003e\n\u003cp\u003eTIDE, Tumor Immune Dysfunction and Exclusion\u003c/p\u003e\n\u003cp\u003eGSEA, gene set enrichment analysis\u003c/p\u003e\n\u003cp\u003eMSigDB, Molecular Signatures Database\u003c/p\u003e\n\u003cp\u003eFDR, false discovery rate\u003c/p\u003e\n\u003cp\u003eBH, Benjamini-Hochberg\u003c/p\u003e\n\u003cp\u003eLASSO, Least Absolute Shrinkage and Selection Operator\u003c/p\u003e\n\u003cp\u003eROC, receiver operating characteristic\u003c/p\u003e\n\u003cp\u003eAUC, area under the curve\u003c/p\u003e\n\u003cp\u003eDCA, decision curve analysis\u003c/p\u003e\n\u003cp\u003ePPI, protein\u0026ndash;protein interaction\u003c/p\u003e\n\u003cp\u003eTF, transcription factor\u003c/p\u003e\n\u003cp\u003emiRNA, microRNA\u003c/p\u003e\n\u003cp\u003emRNA, messenger RNA\u003c/p\u003e\n\u003cp\u003eRBP, RNA-binding protein\u003c/p\u003e\n\u003cp\u003eIL, interleukin\u003c/p\u003e\n\u003cp\u003eNK, natural killer\u003c/p\u003e\n\u003cp\u003eSNP, single nucleotide polymorphism\u003c/p\u003e\n\u003cp\u003eSNV, single nucleotide variant\u003c/p\u003e\n\u003cp\u003eCDF, empirical cumulative distribution function\u003c/p\u003e\n\u003cp\u003ePCA, principal component analysis\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePartial financial support was received from the Youth Development Program of the First Affiliated Hospital of Naval Medical University (No. 2021JCQN04) and General Program Incubation Program of Naval Medical University (No. 2023MS017).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eZixin Liu, Ruijiao Kong and Zichen Liu contributed equally to this work. Zixin Liu, Ruijiao Kong, Yin Jia and Gang Jin were involved in the study design; Ruijiao Kong and Zichen Liu collected and processed the data. Zixin Liu, Ruijiao Kong and Zichen Liu analyzed and interpreted the data. Zixin Liu, Ruijiao Kong and Yin Jia contributed to the analysis methods. Zixin Liu, Ruijiao Kong, Zichen Liu and Gang Jin contributed to the writing of the manuscript. All the authors approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflicts of Interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors have no relevant financial or non-financial interests to disclose.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;and\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003econsent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSince the data adopted in this study were all publicly available data from the TCGA, GEO, GeneCards and DRESIS database, all data related studies were approved by their respective ethical review committees and received written informed consent from patients. Therefore, this study does not need additional ethics approval.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to publish\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData are available on reasonable request. All data relevant to the study are included in the article or uploaded as supplemental information.TCGA_PAAD is downloaded from The Cancer Genome Atlas (TCGA) database (https://www.cancer.gov/ccg/research/genome-sequencing/tcga). Microarray data sets (GSE28735 and GSE71729) are downloaded from the Gene Expression Integrated Database (GEO) (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE28735 and https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE71729).\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eYueting L, Xin S, Shuai L et al. Development and validation of a predictive model based upon extracellular vesicle-derived transposable elements for non-invasive detection of pancreatic adenocarcinoma. 2025;13(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s40364-025-00770-6\u003c/span\u003e\u003cspan address=\"10.1186/s40364-025-00770-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXiaolan H, Zhengyang X, Ruiping R et al. A novel sphingolipid metabolism-related long noncoding RNA signature predicts the prognosis, immune landscape and therapeutic response in pancreatic adenocarcinoma. 2024;10(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.heliyon.2023.e23659\u003c/span\u003e\u003cspan address=\"10.1016/j.heliyon.2023.e23659\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiuyuan X, Xin G, Dong L et al. S100A13-driven interaction between pancreatic adenocarcinoma cells and cancer-associated fibroblasts promotes tumor progression through calcium signaling. 2025;23(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s12964-025-02049-7\u003c/span\u003e\u003cspan address=\"10.1186/s12964-025-02049-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLan THP, Yating C, Yohei F et al. AXL promotes inflammatory breast cancer progression by regulating immunosuppressive macrophage polarization. 2025;27(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s13058-025-02015-8\u003c/span\u003e\u003cspan address=\"10.1186/s13058-025-02015-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChristopher JH, Corbin P, Ilya K et al. Macrophage-Released Pyrimidines Inhibit Gemcitabine Therapy in Pancreatic Cancer. 2019;29(6). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.cmet.2019.02.001\u003c/span\u003e\u003cspan address=\"10.1016/j.cmet.2019.02.001\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHibiki U, Kunitoshi S, Toshiaki T et al. ADAR1-high tumor-associated macrophages induce drug resistance and are therapeutic targets in colorectal cancer. 2025;24(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s12943-025-02312-y\u003c/span\u003e\u003cspan address=\"10.1186/s12943-025-02312-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarrett T, Wilhite SE, Ledoux P, et al. NCBI GEO: archive for functional genomics data sets\u0026ndash;update. Nucleic Acids Res. 2013. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/nar/gks1193\u003c/span\u003e\u003cspan address=\"10.1093/nar/gks1193\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. 41(Database issue):D991-5.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStelzer G, Rosen N, Plaschkes I, et al. Curr protocols Bioinf. 2016;54. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/cpbi.5\u003c/span\u003e\u003cspan address=\"10.1002/cpbi.5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. The GeneCards Suite: From Gene Data Mining to Disease Genome Sequence Analyses1.30.1-1.3.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSun X, Zhang Y, Li H, et al. DRESIS: the first comprehensive landscape of drug resistance information. Nucleic Acids Res. 2023;51(D1):D1263\u0026ndash;75. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/nar/gkac812\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkac812\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYin F, Liu X, Li D, Wang Q, Zhang W, Li L. Tumor suppressor genes associated with drug resistance in ovarian cancer (review). Oncol Rep. 2013;30(1):3\u0026ndash;10. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3892/or.2013.2446\u003c/span\u003e\u003cspan address=\"10.3892/or.2013.2446\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu H, Zhang W, Xu XR, Chen S. Drug resistance related genes in lung adenocarcinoma predict patient prognosis and influence the tumor microenvironment. Sci Rep. 2023;13(1):9682. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41598-023-35743-y\u003c/span\u003e\u003cspan address=\"10.1038/s41598-023-35743-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang J, Li X, Wang K, et al. CLEC7A regulates M2 macrophages to suppress the immune microenvironment and implies poorer prognosis of glioma. Front Immunol. 2024;15:1361351. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fimmu.2024.1361351\u003c/span\u003e\u003cspan address=\"10.3389/fimmu.2024.1361351\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi S, Zhang N, Yang Y, Liu T. Transcriptionally activates CCL28 expression to inhibit M2 polarization of macrophages and prevent immune escape in colorectal cancer cells. Translational Oncol. 2024;40:101842. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.tranon.2023.101842\u003c/span\u003e\u003cspan address=\"10.1016/j.tranon.2023.101842\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang Z, Wang J, Pi J, et al. Identification and Validation of Genes Related to Macrophage Polarization and Cell Death Modes Under Mycobacterium tuberculosis Infection. J Inflamm Res. 2024;17:1397\u0026ndash;411. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2147/jir.S448372\u003c/span\u003e\u003cspan address=\"10.2147/jir.S448372\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHu P, Li B, Yin Z, et al. Multi-omics characterization of macrophage polarization-related features in osteoarthritis based on a machine learning computational framework. Heliyon. 2024;10(9):e30335. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.heliyon.2024.e30335\u003c/span\u003e\u003cspan address=\"10.1016/j.heliyon.2024.e30335\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShi Z, Shen J, Qiu J, Zhao Q, Hua K, Wang H. CXCL10 potentiates immune checkpoint blockade therapy in homologous recombination-deficient tumors. Theranostics. 2021;11(15):7175\u0026ndash;87. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.7150/thno.59056\u003c/span\u003e\u003cspan address=\"10.7150/thno.59056\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBao Y, Jiang A, Dong K, et al. DDX39 as a predictor of clinical prognosis and immune checkpoint therapy efficacy in patients with clear cell renal cell carcinoma. Int J Biol Sci. 2021;17(12):3158\u0026ndash;72. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.7150/ijbs.62553\u003c/span\u003e\u003cspan address=\"10.7150/ijbs.62553\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang P, Gu S, Pan D, et al. Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nat Med. 2018;24(10):1550\u0026ndash;8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41591-018-0136-1\u003c/span\u003e\u003cspan address=\"10.1038/s41591-018-0136-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiberzon A, Subramanian A, Pinchback R, Thorvaldsd\u0026oacute;ttir H, Tamayo P, Mesirov JP. Molecular signatures database (MSigDB) 3.0. Bioinformatics (Oxford. England). 2011;27(12):1739\u0026ndash;40. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/bioinformatics/btr260\u003c/span\u003e\u003cspan address=\"10.1093/bioinformatics/btr260\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSzklarczyk D, Gable AL, Lyon D, et al. STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res. 2019;47(D1):D607\u0026ndash;13. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/nar/gky1131\u003c/span\u003e\u003cspan address=\"10.1093/nar/gky1131\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFranz M, Rodriguez H, Lopes C, et al. GeneMANIA update 2018. Nucleic Acids Res. 2018;46(W1):W60\u0026ndash;4. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/nar/gky311\u003c/span\u003e\u003cspan address=\"10.1093/nar/gky311\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou KR, Liu S, Sun WJ, et al. ChIPBase v2.0: decoding transcriptional regulatory networks of non-coding RNAs and protein-coding genes from ChIP-seq data. Nucleic Acids Res. 2017;45(D1):D43\u0026ndash;50. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/nar/gkw965\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkw965\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi JH, Liu S, Zhou H, Qu LH, Yang JH. Nucleic Acids Res. 2014;42(Database issue):D92\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/nar/gkt1248\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkt1248\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. starBase v2.0: decoding miRNA-ceRNA, miRNA-ncRNA and protein-RNA interaction networks from large-scale CLIP-Seq data.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDeepak R et al. In Hwan P, Timothy V P,. Syk Inhibition Reprograms Tumor-Associated Macrophages and Overcomes Gemcitabine-Induced Immunosuppression in Pancreatic Ductal Adenocarcinoma. 2023;83(16). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1158/0008-5472.Can-22-3645\u003c/span\u003e\u003cspan address=\"10.1158/0008-5472.Can-22-3645\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBen G, Olga K, Andrey K et al. Transcriptomic-Based Microenvironment Classification Reveals Precision Medicine Strategies for Pancreatic Ductal Adenocarcinoma. 2024;166(5). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1053/j.gastro.2024.01.028\u003c/span\u003e\u003cspan address=\"10.1053/j.gastro.2024.01.028\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRebecca LS, Angela NG, Ahmedin JJCCJC, Cancer statistics. 2024. 2024;74(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3322/caac.21820\u003c/span\u003e\u003cspan address=\"10.3322/caac.21820\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDaan JK, Leila AJC. Macrophages at the interface of the co-evolving cancer ecosystem. 2023;186(8). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.cell.2023.02.020\u003c/span\u003e\u003cspan address=\"10.1016/j.cell.2023.02.020\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eN\u0026uacute;ria V-B, Neus M-B, Pablo GF, Pilar NJE. Hallmarks of pancreatic cancer: spotlight on TAM receptors. 2024;107(0). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ebiom.2024.105278\u003c/span\u003e\u003cspan address=\"10.1016/j.ebiom.2024.105278\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHong CE, Lyu SY. Modulation of Breast Cancer Cell Apoptosis and Macrophage Polarization by Mistletoe Lectin in 2D and 3D Models. Int J Mol Sci. 2024;25(15). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/ijms25158459\u003c/span\u003e\u003cspan address=\"10.3390/ijms25158459\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHye-Soo P, Yong Woo B, Yeo-Jin S, Hwa-Jung KJC. Mycobacterium avium subsp. paratuberculosis MAP1889c Protein Induces Maturation of Dendritic Cells and Drives Th2-biased Immune Responses. 2020;9(4). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/cells9040944\u003c/span\u003e\u003cspan address=\"10.3390/cells9040944\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQing L, Chaogang Y, Shuyi W et al. Wnt5a-induced M2 polarization of tumor-associated macrophages via IL-10 promotes colorectal cancer progression. 2020;18(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s12964-020-00557-2\u003c/span\u003e\u003cspan address=\"10.1186/s12964-020-00557-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJennifer S, Stefan C, Feda A et al. IL-10R inhibition reprograms tumor-associated macrophages and reverses drug resistance in multiple myeloma. 2024;38(11). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41375-024-023.91-8\u003c/span\u003e\u003cspan address=\"10.1038/s41375-024-023.91-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThomas FS, Ammar AJ, Atsushi O, et al. Pancreat cancer. 2025;405(10485). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/s0140-6736(25)00261-2\u003c/span\u003e\u003cspan address=\"10.1016/s0140-6736(25)00261-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEun-Sook P, Mehrangiz D, Ah Ram L, Bo Bae P, Kyun-Hwan KJC. Dysregulation of Liver Regeneration by Hepatitis B Virus Infection: Impact on Development of Hepatocellular Carcinoma. 2022;14(15). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/cancers14153566\u003c/span\u003e\u003cspan address=\"10.3390/cancers14153566\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYong L, Xiaolong X, Guoyou Z et al. SOCS2 inhibits hepatoblastoma metastasis via downregulation of the JAK2/STAT5 signal pathway. 2023;13(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41598-023-48591-7\u003c/span\u003e\u003cspan address=\"10.1038/s41598-023-48591-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDan L, Yandan H, Min W, Bin C, Yan LJCCHTS. Overexpression of SOCS2 Inhibits EMT and M2 Macrophage Polarization in Cervical Cancer via IL-6/JAK2/STAT3 Pathway. 2023;27(7). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2174/1386207326666230818092532\u003c/span\u003e\u003cspan address=\"10.2174/1386207326666230818092532\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYuan Z, Jinheng Q, Yanneng F et al. Inhibition of DDR1 promotes ferroptosis and overcomes gefitinib resistance in non-small cell lung cancer. 2024;1870(7). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.bbadis.2024.167447\u003c/span\u003e\u003cspan address=\"10.1016/j.bbadis.2024.167447\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMingjie H, Yi C, Zhiping Y, Lixia Z, Huamei LJABP. Upregulation of SOCS2 causes mitochondrial dysfunction and promotes ferroptosis in pancreatic cancer cells. 2023;70(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.18388/abp.2020_6383\u003c/span\u003e\u003cspan address=\"10.18388/abp.2020_6383\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"pancreatic adenocarcinoma, drug resistance, macrophage polarization, molecular characterization, immune microenvironment","lastPublishedDoi":"10.21203/rs.3.rs-7991235/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7991235/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackgroud: \u003c/strong\u003ePancreatic adenocarcinoma (PAAD) is characterized by an aggressive behavior and poor prognosis, requiring innovative therapeutic strategies.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e The PAAD datasets were acquired from two publicly available genomic repositories: The Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus (GEO) database. Drug resistance- and macrophage polarization-related genes (DMRGs) were collected on GeneCards or DRESIS databases. To identify distinct disease subtypes, we identifiedprognostic genes with univarite COX regression analysis, followed by consensus clustering. Then an intersection analysis between differentially expressed genes (DEGs) and a set of DMRGs was performed, and the overlapping genes yielded drug resistance- and macrophage polarization-related differentially expressed genes (DMRDEGs). Based on DMRDEGs identified, a prognostic risk model was constructed.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e PAAD patients were categorized into two molecularly distinct subgroups, subtype A (1) and subtype B (2), based on DMRGs. Through immunological profiling, we found five distinct immune cell populations with statistically significant variations, notably comprising regulatory T lymphocytes and activated NK cells. Immunological profiling demonstrated that subtype B displayed increased sensitivity to immunotherapy (p \u0026lt; 0.01). A prognostic risk model comprising five key genes (\u003cem\u003eIL18, EREG, LDHA, SOCS2, \u003c/em\u003eand\u003cem\u003eSPP1\u003c/em\u003e) was built and showed robust predictive capability (area under the curve (AUC) \u0026gt; 0.7). A protein-protein interaction (PPI) network was established focusing on these genes, revealing their function as key regulatory hubs.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion:\u003c/strong\u003e\u003cem\u003e\u003cstrong\u003e \u003c/strong\u003e\u003c/em\u003eOur analysis categorized PAAD into two distinct subgroups based on DMRGs and a prognostic risk model developed from these genes exhibits considerable promise for forecasting patient survival outcomes.\u003c/p\u003e","manuscriptTitle":"Novel drug resistance- and macrophage polarization-related molecular subtyping and prognostic signature for pancreatic adenocarcinoma","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-22 08:50:27","doi":"10.21203/rs.3.rs-7991235/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"3cef849d-87bf-400c-8396-5c9f55cd0053","owner":[],"postedDate":"December 22nd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-02-23T19:54:54+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-22 08:50:27","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7991235","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7991235","identity":"rs-7991235","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}