Proteomic Atlas of Post-COVID Sequelae

Proteomic Atlas of Post-COVID Sequelae | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Proteomic Atlas of Post-COVID Sequelae Philipp Wild, Sepehr Golriz Khatami, Rieke Baumkötter, Thomas Koeck, and 13 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6024310/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract The emergence of post-COVID sequelae (PCS) represents a global challenge. However, understanding of biological mechanisms and the definition of quantifiable risk factors remains limited. This study harnessed the power of predictive machine learning models to explore the potential of proteomics in predicting individual post-COVID symptoms and their collective manifestation as PCS. The analysis utilized a panel of approximately 2900 proteins measured in 495 COVID-19 patients. The study identified 235 unique proteins associated with 21 distinct post-COVID symptoms. Symptoms more closely linked by similar protein patterns tended to co-occur more frequently in patients. Six symptom clusters with distinct molecular pathway signatures were uncovered, with metabolic and inflammatory pathways prominently involved across several clusters. The relevance of the specific protein signatures for post-COVID symptoms could be demonstrated and explored by objective, quantifiable clinical tests, including cognitive and somatic assessments, and underlined their relevance. Data from various modalities, including pre-existing conditions, disease risk factors and genetic susceptibility, revealed relevant relations that may contribute to PCS heterogeneity. This work underscores the complex and multifaceted nature of post-COVID symptoms. It emphasizes the need for systematic and more specific approaches to facilitate the development of targeted therapies and treatment strategies. Health sciences/Diseases/Infectious diseases/Viral infection Biological sciences/Computational biology and bioinformatics/Protein function predictions Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 INTRODUCTION Amid the ongoing global SARS-COV-2 infections, the emergence of post-COVID sequelae (PCS), affecting 30–50% of COVID-19 survivors, presents novel challenges.[ 1 , 2 ] PCS is a syndrome characterized by a wide range of symptoms, from more common ones, including fatigue, to more severe manifestations such as neurological disorders.[ 3 ] PCS is defined by the World Health Organization (WHO) as the emergence of new symptoms or the persistence of symptoms beyond three months following a SARS-COV-2 infection in the absence of an alternative explanation.[ 4 ] Various factors such as demographic attributes (e.g., age and sex), pre-existing somatic and cognitive conditions, and history of hospitalization for SARS-COV-2, including admission to an intensive care unit, are linked with an increased risk of PCS.[ 5 ] Several putative pathophysiologic mechanisms, namely viral persistence, hypercoagulopathy, immune dysregulation, autoimmunity, hyperinflammation, imbalance of the renin-angiotensin system, and oxidative stress, have been proposed as causing PCS.[ 6 , 7 ] The advent of high-throughput proteomic assays enabled the measurement of thousands of proteins, facilitating the systematic examination of molecular changes throughout the disease and providing comprehensive insight into the biological mechanisms and disease etiology.[ 8 ] Proteomics has been widely employed in SARS-COV-2 research.[ 9 – 13 ] For instance, an increase in protein expression level related to neutrophils or the complement system was found in severe disease courses.[ 9 , 12 ] Proteins regulating the circadian rhythm pathway were identified in higher concentrations in post-COVID patients with neurological abnormalities [ 14 ]. Proteomics was also employed to develop a classification model to predict long COVID using proteomic data from acute-phase SARS-COV-2 patients with 12-month symptom persistence.[ 10 ] Although these studies shed light on some aspects of the syndrome, their reach and impact are limited as brief follow-up periods, small cohort sizes, suboptimal protein coverage, and a lack of comprehensive long-term data hindered the ability to draw definitive conclusions. This study leveraged a high-throughput proteomic technology by proximity extension assay technology (Olink, Uppsala, Sweden) to generate comprehensive protein profiles of SARS-COV-2 survivors from plasma samples of a population-based cohort. The aim was to employ machine-learning models to identify protein signatures associated with developing of post-COVID symptoms. Symptoms-associated proteins were investigated concerning objective clinical tests, encompassing cognitive, psychiatric, and somatic assessments, to explore the potential clinical utility of the detected proteins. The work also capitalized on the synergy between multi-automated data and comprehensive phenotypic information to characterize key biological factors associated with PCS heterogeneity. RESULTS Sample characteristics The clinical characteristics of 495 SARS-COV-2 survivors from GCS included in this study, with a mean age of 53.2±15.8 years and an approximately balanced sex ratio (47.9% female) are shown in Table 1 . Among cardiovascular risk factors, arterial hypertension emerged as the most prevalent, affecting 43.8% of the study sample whereas type 2 diabetes mellitus and smoking were observed with lower frequencies. Similarly, pre-existing cardiovascular disease was the most common comorbidity, with a prevalence of 12.7%. Chronic liver disease and atrial fibrillation (AF) were less frequent, affecting 2.0% and 2.2% of participants, respectively. Symptoms of post-COVID and machine learning-derived protein signatures Study participants were asked about 61 distinct symptoms (see Supplementary Methods ). The symptom catalogue was used for the development of protein-based models explaining the presence of post-COVID or specifically symptoms. All symptoms reported by fewer than 10 individuals were excluded from subsequent analysis to ensure the reliability and robustness of predictive models. Overall this led to 21 symptoms to be further explored. Fatigue was the most prevalent symptom, affecting 11.7% of individuals (N=58), while slow movements (2.0%; n=10) and depression (2.2%; N=11) were less frequent manifestations within the sample. Upon completion of the quality control, 46 proteins from an initial set of 2,945 were excluded due to being below the limit of detection, low call rates, or technical variability. Of 2,899 remaining proteins, 235 unique ones were associated with 15 distinct post-COVID symptoms. The model for predicting the presence of post-COVID regardless of the symptoms present, exhibited a lower discriminative ability (leave-one-out cross validation; LOOCV AUC = 0.62) compared to the symptom-specific predictive models. Among the models, each tailored to a specific symptom, the one for predicting chest pain as a long-term sequela of SARS-CoV-2 exhibited the highest LOOCV AUC of 0.84, while the one capturing anxiety showed the lowest discriminative ability with an AUC of 0.51. The predictive models could not differentiate between individuals who exhibited the following symptoms and those who did not, across six distinct symptoms namely balance issues, cough, joint pain, and muscle pain, headache, and sleep disturbances ( Figure 1 ). The ‘shortness of breath’ symptom had the highest number of associated proteins, totaling N=51. Conversely, the symptoms ‘concentration problems’ with counts of five proteins including protein kinase AMP-activated non-catalytic subunit gamma 3 (PRKAG3), Ketohexokinase (KHK), Calcineurin like EF-hand protein 1 (CHP1), Cell division cycle 25A (CDC25A), Annexin A1 (ANXA1) and ‘anxiety’ with three (Interleukin 1 alpha; IL1A, ANXA1, Switching B cell complex subunit; SWAP70) had the fewest proteins. Interestingly, some proteins were associated with several symptoms, the connection between which also showed clinical plausibility ( Figure 2 ): IL1A is the most frequently detected protein, associated with five distinct symptoms, including shortness of breath, hair loss, depression, anxiety, and memory problems. Disturbances in olfaction and taste were found to have the highest number of shared proteins (N=11), with Toll like receptor 4 (TLR4), Thyroglobulin (TG), Melanotransferrin (MELFT), Creatine kinase, mitochondrial 1A (CKMT1A), Chymotrypsin like (CTRL), Eukaryotic translation initiation factor 1A X-linked (EIF1AX), Apolipoprotein B receptor (APOBR), Sphingomyelin phosphodiesterase acid like 3B (SMPDL3B), Asporin (ASPN), and TSPY like 1 (TSPYl1), followed by shortness of breath and fatigue (CD3 gamma subunit of T-cell receptor complex; CD3G, Pancreatic lipase related protein 2; PNLIPRP2, Sulfatase modifying factor 2; SUMF2, Endosome associated trafficking regulator 1; ENTR1) as well as shortness of breath and depression (MANSC domain containing 4; MANSC4, IL1A, S100 calcium binding protein A11; S100A11, and Paraoxonase 3; PON3) each sharing four proteins. To elucidate the relationship between the symptom-specific protein signatures a correlation matrix was constructed ( Figure 3A ): The strongest positive correlations were observed for slow movements and loss of interest (r=0.82) and for anxiety and palpitations (r=0.75). In contrast, the most pronounced inverse correlations were identified for mood swings and olfactory disturbances (r=-0.21) as well as for concentration problems and olfactory disturbances (r=-0.19). Furthermore, the analysis revealed higher positive correlations between fatigue and slow movements (r=0.57), concentration and memory problem (r=0.53) and olfactory and taste disturbances (r=0.52). Subsequently, the co-prevalence of clinical post-COVID symptoms was compared with the correlation of the respective symptom-related protein scores and showed a moderate positive correlation of 0.36 ( Figure 3B ). Putative pathomechanisms underlying post-COVID related symptoms To gain a deeper understanding of the biological mechanisms underlying post-COVID symptoms, first, hierarchical clustering was employed to group symptoms based on the correlation of their corresponding protein signatures. Proteins associated with the co-clustered symptoms were then subjected to pathway enrichment analysis. The clustering analysis yielded six discrete symptom clusters, as illustrated in Figure 3C: cluster 1 with fatigue/tiredness, shortness of breath, loss of interest or pleasure, and slow movements; cluster 2 with anxiety, and palpitations; cluster 3 with hair loss, memory problems, mood swings, and concentration problems; cluster 4 with olfactory disturbances, and taste disturbances; cluster 5 with sleep disturbances, and depression; and cluster 6 with chest pain. A total of 59 distinct molecular pathways were identified as relevantly associated with symptom clusters. The largest number of molecular pathways were identified in symptom cluster 1 (n = 20), while the smallest number were identified in symptom cluster 6 (n = 1). The most prevalent class of pathways within symptom cluster 1 were those pertaining to metabolism, including fatty acid metabolism and the regulation of lipid transport. Additionally, pathways associated with signalling, such as those involving ATM signalling and P53 downstream, were also highly represented. The most common pathways associated with symptom cluster 2 were those related to cellular stress and damage response, including pathways involved in the response to growth factors and temperature stimuli. In contrast, metabolic pathways were the most frequent category of pathways associated with symptom cluster 4. The most frequently occurring pathways associated with symptom cluster 3 were those pertaining to the immune and inflammatory responses and those involved in signalling. The sole pathway found to be associated with symptom cluster 6 is a cell proliferation and regulation pathway (i.e., positive regulation of exocytosis). While the enrichment analysis revealed distinct pathways associated with individual clusters, several pathways were identified as shared across multiple symptom clusters, notably the "response to temperature stimulus" (symptom clusters 1, 2, 3), "carbohydrate derivative catabolic process" (symptom clusters 1, 4), and "response to wounding" (symptom clusters 2, 3). Clinical validation of post-COVID symptom-specific protein signatures To further explore the potential clinical utility of proteins discovered for individual post-COVID symptoms, their link with measurements from various established and objective clinical tests, encompassing cognitive, psychiatric, and somatic assessments were analyzed ( Figure 4 ). Broadly, a correlative relationship between symptom-centered objective clinical tests and the protein score associated with the corresponding symptom was detected. For example, a positive association was found between protein signature related with mood swings and the nine-item Patient Health Questionnaire (PHQ-9) [15] , a depressive symptom scale. Similarly, the sleep disturbances- protein signature was associated with the Jenkins Sleep Scale (JSS) [16] , a scale for estimating of sleep problems in clinical research. In contrast, a negative association was identified between the Montreal Cognitive Assessment (MoCA) [17] as a screening measure for the diagnosis of cognitive impairment and the protein signatures corresponding to cognitive function, specifically memory with regression coefficients (β = -0.11) and concentration with (β ≈ -0.1). Similarly, an analogous association was detected between the objective clinical olfactory test (i.e., a smell test comprising 12 smelling sticks) and the protein signatures indicative of olfactory and gustatory modalities, individually. However, no clinically relevant link was found between spirometry tests measuring vital capacity (VC) and forced expiratory volume in one second (FEV1) and the protein signature for shortness of breath. Relation of the clinical patient profile with post-COVID symptom-specific protein signatures The association between pre-existing conditions and protein signatures of individual symptoms was investigated to characterize the clinical factors, like risk factors and comorbidities, that may be predisposing to or even be involved in the development of post-COVID symptoms, the association between pre-existing conditions and the protein signatures of individual symptoms was investigated ( Figure 5 ). The analyses revealed the strongest positive association between diabetes mellitus (DM) and the protein signature associated with memory problems (β per SD = 0.28, 90%Cl [-0.05; 0.6]), while the strongest inverse association was detected between obesity and the protein signature for hair loss (β per SD = -0.28, 95%Cl [-0.54; -0.03]). Positive associations were observed between smoking and protein signatures linked to taste (β per SD = 0.21, 90%Cl [0; 0.41]) and olfactory disturbances (β per SD = 0.16, 90%Cl [-0.02; 0.35]). Furthermore, the study highlighted the influence of comorbidities on the prevalence of diverse post-COVID symptoms, as indicated by the protein signatures. The protein signature of sleep disturbances exhibited an inverse association with the largest number of comorbidities including coronary artery disease (CAD), peripheral arterial disease, history of pulmonary embolism, congestive heart failure (CHF), and arterial fibrillation (AF). The strongest association was observed for CHF (β = 1.38, 95%Cl 0.34; 2.43]). In contrast, the protein signatures of the post-COVID symptoms namely anxiety, memory problems, palpitations, olfactory disturbances, chest pain, loss of interest and slow movements were only associated with one comorbidity each. These were a history of venous thromboembolism (VTE), CAD, AF, history of stroke and chronic obstructive pulmonary disease (COPD). Furthermore, the investigation demonstrated that the strongest positive standardized association was observed for the relation of H.x VTE with the protein signature of shortness of breath (β per SD =1.48, 95%Cl [-1.82; 0.64]), while the strongest inverse association was noted for COPD with the signature of loss of interest (β per SD =-1.3, 90%Cl [-0.35; 1.46]). Genetic predisposition for cardiovascular risk factors and diseases and susceptibility of post-COVID symptoms Finally, the potential impact of a genetic predisposition for cardiovascular risk factors and diseases on the susceptibility to post-COVID symptoms was evaluated using polygenic risk scores and the protein signatures of post-COVID symptoms ( Figure 6 ). Several associations were identified between polygenic risk scores and various post-COVID manifestations: The most substantial positive association was observed between the polygenic risk scores for cholesterol concentration and the "loss of interest" proteins (β = 0.36). Conversely, the highest negative associations were found between the hyperlipidaemia-computed polygenic risk score and the protein signatures for "slow movement" (β = -0.23) and for “memory problem" (β = -0.21). A positive correlation was detected between the polygenic risk score for stroke and the protein signature for chest pain (β = 0.25) as well as between the cholesterol-estimated polygenic risk score and the proteins linked to fatigue (β = 0.20). Conversely, an inverse association was observed between the protein signature for olfactory disturbance and the polygenic risk score for smoking (β = -0.1). DISCUSSION This study generated protein signatures of the 15 most common symptoms of the post-COVID, investigated their interrelations, and explored the putative underlying pathomechanisms. The analysis of associations between the symptom-specific protein signatures and objective clinical tests demonstrated the potential clinical utility of detected proteins. The study also examined the relevance of pre-existing conditions and disease risk factors with symptom-related protein scores, and the link between genetic predisposition of clinical traits and disease susceptibility with protein scores of individual symptoms. Overall, machine learning models demonstrated superior discriminative capabilities concerning individual symptoms, as opposed to the protein profile associated with the general presence of a post-COVID. A more comprehensive understanding may therefore be achieved by deconstructing the syndrome into distinct symptom clusters, each potentially linked to different molecular mechanisms. Concerning protein-symptom associations, Interleukin-1 alpha (IL1A) emerged as the most frequently identified protein. It was associated with five distinct clinical symptoms: dyspnea, hair loss, depression, anxiety, and cognitive impairments. IL1A, a known proinflammatory mediator, crucial role modulating inflammation and SARS-COV-2 pathogenesis.[ 18 ] Studies have shown that SARS-COV-2 infection increases interleukin-1 (IL-1) expression leading to inflammatory factor accumulation in the lungs and potentially triggering a cytokine storm and chronic inflammation.[ 19 ] This persistent inflammatory condition may be associated with immunothrombosis and the formation of microclots, impede oxygen exchange, and contribute to shortness of breath. Similarly, inflammatory cytokines such as IL-1 can disrupt the hair growth cycle, causing increased shedding and hair loss [ 20 ] and hinder hair regrowth.[ 21 ] Additionally, Paraoxonase 3 (PON3) was also associated with a cluster of symptoms, including fatigue, slow movements, depression, mood swings, and loss of interest. The extensive inflammation observed during SARS-COV-2 infection dysregulates the expression of the paraoxonase family (PON1, PON2, PON3)[ 22 ]. This dysregulation can lead to increased oxidative stress, as PON enzymes play a crucial role in protecting against it by detoxifying lipid peroxides and preventing low-density lipoprotein (LDL) oxidatio [ 23 ], in which elevated LDL cholesterol has been associated with symptoms such as fatigue, depression, and diminished interest. [ 24 ] Several symptoms were co-prevalent, including the well-known co-prevalence of olfactory and taste disturbances. Various studies reported an intimate functioning between the chemosensory systems, specifically smell and taste, and about 30% of SARS-COV-2-infected individuals experienced a dysfunction of both systems.[ 25 ] Contrary to previous studies [ 26 ], but in concordance with Park et al. [ 27 ], an inverse relationship between olfactory function and concentration problems was observed. The correlation identified between mental and somatic symptoms, particularly anxiety and palpitations, is supported by studies showing that patients with affective disorders often have symptoms of autonomic vascular dystonia, including severe palpitations. [ 28 ] This also relates to the link between concentration problems, memory problems, and mood, as depression often leads to difficulties in remembering information, especially positive information [ 29 ]. The pathway enrichment analysis revealed a wide range of pathway categories from metabolic and signaling pathways to cellular stress and damage response pathways, as well as immune and inflammatory pathways. Among pathways associated with symptoms in cluster 1, the ATM signaling pathway, signaling by interleukins, fatty acid metabolism, and apoptotic signaling pathway were found, all of which play roles in the development fatigue/tiredness [ 30 ], shortness of breath [ 31 ], loss of interest and pleasure [ 32 ], or slow movements [ 33 ]. Different mechanisms have been proposed for how apoptotic signaling pathway contributes to the development and manifestation of post-COVID, namely persistent inflammation and tissue damage, dysregulated immune responses, autoimmunity, and mitochondrial dysfunction. [ 34 ] The regulation of leukocyte migration and the regulation of kinase activity are two of the pathways linked to symptoms in cluster 3. Studies have shown that SARS-COV-2 is associated with abnormal leukocyte migration.[ 35 ] Recently lymphatic vessels in the membranes around the brain were discovered and the lymphatic system was introduced as a potential new player in complex neurological problems such as memory impairment.[ 36 ] The analysis of cluster 6 resulted in the regulation of exocytosis as a pathway associated with the sole symptom chest pain in this cluster. The known endothelial cell inflammation (endotheliitis) in SARS-COV-2 [ 37 ] was reported to lead to degranulation and exocytosis of Weibel-Palade bodies containing von Willebrand factor, promoting recruitment and aggregation of platelets[ 38 ], which in turn, may increase the risk of thrombosis and subsequent chest pain, specifically in the case of heart or lung affection. [ 39 ] The co-existence of diverse symptoms within a single cluster may hint at shared pathophysiological mechanisms across these symptoms. This is supported by the pathway enrichment analysis, which revealed common pathways enriched within each cluster and pathways shared across distinct clusters. For instance, the response to temperature stimulus of clusters 1, 2, and 3 explains the dysregulation of temperature response pathways, mediated through neuroinflammation, cytokine signaling, and autonomic dysfunction. [ 40 ] The association between protein-specific molecular signatures and quantitative clinical assessments underscores their diagnostic and therapeutic potential. Several proteins have been identified as being associated with olfactory dysfunction in post-COVID-19 symptoms, including Toll-Like Receptor 4 (TLR4), a mediator of inflammatory responses within the olfactory epithelium; Annexin A1 (ANXA1), an anti-inflammatory mediator; Surfactant Protein A1 (SFTPA1), a component of mucosal immunity; Ring Finger Protein 5 (RNF5), a regulator of protein quality control; Prostaglandin E Synthase 2 (PTGES2), involved in inflammatory pathways; and Nudix Hydrolase 15 (NUDT15), a metabolic regulator. These proteins have been investigated as potential therapeutic targets for olfactory disorders. [ 41 – 43 ] Similarly, the analysis identified proteins associated with memory impairment, including Microtubule-Associated Protein Tau (MAPT), a critical contributor to neurodegenerative pathologies; Interleukin 1 Alpha (IL1A), a mediator of neuroinflammation; Neurotrophin-3 (NTF3), essential for synaptic plasticity; Sex Hormone-Binding Globulin (SHBG), which influences cognitive function; Allograft Inflammatory Factor 1 (AIF1), associated with neuroinflammatory responses; and Reticulon 4 Receptor (RTN4R), a regulator of synaptic remodeling. These proteins represent promising therapeutic targets for the treatment of memory dysfunction, as supported by prior research. [ 44 – 46 ] This study also displayed associations between risk factors that may be involved in developing post-COVID symptoms, such as smoking and taste and smell disturbances. Approximately 5% of individuals who report initial chemosensory dysfunction, specifically taste and smell, may experience persistent symptoms six to twelve months following the infection.[ 47 ] The impact of SARS-COV-2 on the sensory functions, could be intensified by the known influence of smoking on the olfactory and gustatory senses [ 48 ]. Further interesting links were observed between comorbidities and post-COVID symptoms, including VTE and chest pain, but also AF and palpitation. VTE-induced inflammation with VTE as both an initiator and a perpetrator of inflammation [ 49 ], may contribute to acute chest discomfort, potentially arising from pleurisy [ 50 ]. Individuals with AF and hypertension exhibited a two-fold increased risk of developing long-term cardiac complications following a SARS-COV-2 infection [ 51 ], potentially induced by inflammatory response, dysfunction of the autonomic nervous system, and electrolyte disturbances. These could trigger electrical and structural remodeling of the atria, thereby intensifying SARS-COV-2-induced palpitations, particularly in individuals with AF. [ 52 ] Interesting associations were detected between genetics by polygenetic risk scores and post-COVID sequelae: Notably, a genetic predisposition for arterial hypertension was associated with the protein signature indicative of concentration problems. While the precise mechanisms between hypertension and cognitive impairment remain elusive, it is postulated that hypertension could precipitate cognitive decline through alterations in blood vessels or brain damage related to vascular complications, inflammation, and oxidative stress.[ 53 ] The link between the genetic susceptibility for higher cholesterol concentrations and the protein pattern for fatigue is supported by earlier reports that suggest that a genetic predisposition to hypercholesterolemia may have an increased vulnerability to post-COVID symptome, particularly fatigue.[ 54 ] The current study has several strengths, including the larger, well-characterized cohort with a longer observation period including a pre-pandemic phase, the measurements with proximity extension technology with over 3000 proteins (Olink explore, Uppsala, Sweden) per individual, and the availability of genetic data. However, some relevant limitations must also be taken into account: First, the work focused on quantifying protein concentrations without considering the conformational state of the proteins. This encompasses their folding patterns and three-dimensional structures and is thus relevant for their functionality and potential interactions within the biological system. Second, the study did not analyze protein dynamics, which limits the conclusions that can be drawn. Third, no replication could be implemented in this exploratory study due to the unique setting of this biobank. Fourth, this study was conducted during the dominance of the wild-type and Alpha variants of SARS-CoV-2 and the results may only be valid for these specific variants. Fifth, while the study encompasses several hundred participants, the subset of individuals reporting a specific post-COVID symptom is still small, limiting the power to identify robust predictors or specific classifications. Finally, mechanistic findings, such as chronic inflammation, may be attributed to underlying pathologies or comorbidities that are not directly relevant to post-COVID. The exclusion of these confounding variables is again limited by the subgroups' relatively small size. This comprehensive study indicates that PCS represents a condition with diverse and yet unexplored underlying pathophysiological mechanisms. The findings shed light on the complex nature of the syndrome, highlighting that factors such as genetic predisposition, medical history, and disease risk factors play a significant role in the observed heterogeneity of the condition. The protein patterns indicate a more comprehensive understanding could be achieved by deconstructing the syndrome into distinct symptom clusters, each linked to different molecular mechanisms. This approach could potentially facilitate the development of targeted therapies and personalized treatment strategies. MATERIALS AND METHODS Study design and population This study mainly leveraged post- COVID data from two large-scale, prospective, population-based cohorts, namely the Gutenberg COVID-19 Study (GCS) and the Gutenberg Post COVID Study (GPC) and background information from Gutenberg Health Study (GHS) [ 15 ]. The individual cohorts provide complementary information that was fused to provide more detailed information on the subjects. This analysis employed i) details on post-infection protein profiles and post-COVID symptoms contributed by GCS, ii) deep (sub)clinical phenotypes measured by objective clinical tests in the course of medical-technical examinations offered by GPC for individuals who have contracted SARS-COV-2 infection, and iii) the wealth of baseline information (i.e., comorbidities and risk factors before the pandemic and genetic data provided by GHS. A detailed description of ethics approval and consent to participate is available in the supplementary information. Gutenberg-COVID19 Study GCS was initiated in 2020 to explore the epidemiological dynamics and population-level impacts of the pandemic across both infected and uninfected population. The GCS enrolled 10,250 participants, 8,121 from GHS aged 45–88 and 2,129 newly recruited from a random younger sample aged 25–44. The cohort maintained a balanced sex distribution, with 50.8% of participants being female. Comprehensive biospecimen collection, including various blood fractions (serum, EDTA plasma, PBMC), stool, tooth pocket swabs, and tear fluid samples and routine laboratory parameters and detailed clinical information (including medical history and family history) were obtained from GCS participants at two time points, approximately four months apart. SARS-CoV-2 infection status was determined using a combination of diagnostic methods: i ) quantitative reverse transcription polymerase chain reaction (RT-qPCR), ii) serological testing for anti-SARS-COV-2 antibodies using two immunoassays covering different antibody spectra i.e., the Abbott Architect SARS-COV-2 IgG (Germany) assay and the Roche Elecsys® Anti-SARS-COV-2 (Germany) assay, and iii) self-reported SARS-CoV-2 infection based on computer-assisted personal interview and weekly smartphone-based assessments in the follow-up period. Subsequently, individuals who were SARS-COV-2-positive and a control group without infection that was matched for demographic characteristics (i.e., age and sex) [ 55 ] were subjected to a computer-assisted telephone interview, asking about post-COVID symptoms and the intensity: Information on 61 symptoms was collected based on the WHO Global SARS-COV-2 Clinical Platform Case Report Form for Post COVID condition. [ 56 ] Participants were requested to report the symptoms (if any) and the duration in three time frames, 0–3 months, 3–6 months, or more than 6 months after infection. They were then classified as having post-COVID if they reported either (1) any new symptom onset after SARS-COV-2 infection, or (2) the persistence of pre-existing symptoms with increased severity after infection. Gutenberg Health Study The Gutenberg COVID-19 study was based on the population sample of the Gutenberg Health Study (GHS), which provided additional information on the health of the participants before the pandemic. Established in 2007, GHS as a population-based prospective cohort study investigates diseases with high social impact, including cardiovascular disease, cancer, ophthalmological diseases, and immune and mental health conditions. With balanced sex representation (1:1 balanced), GHS enrolled over 20,000 participants aged 25–85. On a quinquennial basis at a dedicated study center, extensive data acquisition encompasses parameters for intermediate clinical phenotypes and subclinical disease markers based on medical-technical examinations; medical, environmental, psychosocial, and lifestyle factors; biomaterial including but not limited to serum, Ethylenediaminetetraacetic acid (EDTA) plasma, and peripheral blood mononuclear cells (PBMC), urine, stool, and tear liquid; and genetic information. At the midpoint of the five-year interval, a computer-assisted telephone interview is carried out to survey the investigated diseases. The detailed methodologies employed for the data compilation were documented elsewhere [ 57 ]. Gutenberg Post COVID Study Set up in 2021, the GPC was designed to investigate the long-term effects of SARS-COV-2 infection. The cohort incorporated 607 participants at the time of analysis, aged 20–92 years, ensuring balanced sex representation (1:1). Participants underwent a comprehensive multidisciplinary clinical examination program over several days after the final diagnosis of post-COVID. This investigation included biobanking of various biospecimens for multi-omics analyses: blood fractions (DNA, RNA, serum, EDTA plasma, citrate plasma, and PBMCs), tooth pocket swabs, urine, tear liquid, stool samples, and routine laboratory. Throughout the examinations, a wide range of (sub)clinical phenotypes including but not limited to heart, lungs, brain, and abdomen was evaluated through diagnostic tools such as Holter electrocardiography, spirometry, smell testing, neuroimaging, and cognitive/psychological tests. High-throughput proteomic profiling The venous blood EDTA plasma samples were collected from GCS participants, aliquoted, and stored at -80°C within a dedicated biobank. They were then subjected to multiplex proximity extension assays (PEA), a state-of-the-art technology provided by Olink Bioscience (Uppsala, Sweden) using eight 384-plex panels. This strategic selection enabled the relative quantification of a repertoire of 2,945 unique proteins per individual. A detailed description of the Olink Explore 3072 methodology is provided in the Supplementary Methods . The measurements were reported in Normalized Protein eXpression (NPX) units on a log2 scale, where higher values indicate increased protein concentrations. The NPX data were then transformed (via comparison of identity, logarithm, square root, and negative inverse) to reduce the absolute value of the skewness of each protein. Finally, these transformed data were then winsorized to a radius of 6 absolute median deviations around the median. Data operationalization and statistical analysis Polygenic risk scores, which represent the aggregated effect of risk alleles for polymorphisms at multiple genomic loci, serves as a quantitative measure for assessing genetic susceptibility to multifactorial traits and disease [ 58 ]. In this work, polygenic risk score were used as a proxy for investigating the association between defined genetic predispositions and the risk of PCS. Polygenic risk score was derived using the established equation: where the PRS for an individual (i) is a summation of the Genome-Wide Association Study coefficient (GWAS) estimates for each variant (j), denoted as (β j ), each multiplied by the count of risk alleles present for each respective variant (G ij ). Additionally, \(\:\rho\:\) j is the GWAS p-value for a variant (j) and \(\:\rho\:\) T is the predetermined threshold. A detailed description of PRS estimation is provided in the Supplementary Methods . Demographic and clinical characteristics of the study sample are presented using absolute counts and relative frequencies. The least absolute shrinkage and selection operator (LASSO)-regularized logistic regression model, with leave-one-out cross-validation (LOOCV), was employed to identify a core yet optimal set of proteins without redundant information. Correlation between symptoms was obtained by applying Pearson correlation to the symptom-specific protein score (P s ) . The P s was established as follows: where n is the number of selected proteins by the model, \(\:\varvec{x}\) is the measured value of the protein, and \(\:\beta\:\) is the protein-related coefficient. The relationship between symptom specific protein signature correlations and their co-prevalence was investigated using correlation analysis. Linear regression was utilized to investigate the association between pre-existing conditions and disease risk factors with symptom-specific protein scores. Similarly, linear regression was applied to investigate the relevance of symptom-oriented protein scores and objective clinical tests. The association between polygenetic risk scores and symptom manifestations was analyzed using linear regression. All analyses were adjusted for age and sex and were of exploratory nature with p-values (P) considered as continuous measure of statistical evidence. All statistical analyses were performed in R (version 4.3.1) ( https://www.r-project.org/ ). Bioinformatic analysis To augment the understanding of the results procured, bioinformatics-based methods were employed. Pathway enrichment analysis was conducted using Metascape (v3.5.20240901) [ 59 ] to elucidate the pathways associated with symptom clusters. The igraph R package (version 2.1.4) [ 60 ] was utilized to visually depict the link between the pathways and the symptom clusters. Declarations Conflict of interest There are no conflicts of interest reported regarding this work. Outside the submitted work, PSW reports grants from Bayer AG, non-financial grants from Philips Medical Systems, grants and consulting fees from Boehringer Ingelheim, grants and consulting fees from Novartis Pharma, grants and consulting fees from Sanofi-Aventis, grants, consulting and lecturing fees from Bayer Health Care, grants and consulting fees from Daiichi Sankyo Europe, lecturing fees from Pfizer Pharma, lecturing fees from Bristol Myers Squibb, consulting fees from Astra Zeneca, consulting fees and non-financial support from Diasorin and non-financial support from IEM. PSW is funded by the Federal Ministry of Education and Research (BMBF 01EO1503), Ministry of Science and Health of the State of Rhineland Palatinate (MWG RLP 724 − 0010#2021/0030-1501), and the Federal Institute for Occupational Safety and Health (BAuA, F2447 / 537727) outside the present work. PSW is principal investigator of the German Center for Cardiovascular Research (DZHK) and principal investigator of the DIASyM research core (BMBF DIASyM research core (BMBF 161L0217A, 031L0217A). AKS received financial support by Abbvie, Bayer, Heidelberg Engineering, Novartis, and Santen outside the topic of this work. The remaining authors report no conflict of interest. Funding The study was funded by the European Regional Development Fund and the Ministry of Science and Health of the State of Rhineland-Palatinate (EFRE/REACT-EU, Grant No. 84007232 and No. 84009735); by the Federal Ministry of Education and Research (“EPIC-AI”, Grant No. 01EQ2401A) for the topic of evaluation of new approaches to data analysis and data sharing in long/post COVID-19 research; by the ReALity Initiative of the Life Sciences of the Johannes Gutenberg University Mainz for the establishment of a cell bank; and by the National University Medicine Research Network on Covid-19 (”NaFoUniMedCovid19”, Grant No. 01KX2021) B-FAST for the topics “poor living conditions” and “working conditions” and their association with COVID-19 in the population. Biomaterials were stored at the BioBank Mainz of the University Medical Center Mainz. Acknowledgement We are indebted to all study participants of the Gutenberg COVID-19 study for their contribution. We thank all study staff and researchers involved in the planning and conduct of the study. References Groff D et al (2021) Short-term and long-term rates of postacute sequelae of SARS-CoV-2 infection: a systematic review. JAMA Netw open 4:e2128568–e2128568 Fernandez-de-Las-Peñas C et al (2023) Persistence of post-COVID symptoms in the general population two years after SARS-CoV-2 infection: A systematic review and meta-analysis. J Infect Oronsky B et al (2023) A review of persistent post-COVID syndrome (PPCS). Clin Rev Allergy Immunol 64(1):66–74 Soriano JB et al (2022) A clinical case definition of post-SARS-COV-2 condition by a Delphi consensus. Lancet Infect Dis 22(4):e102–e107 Tsampasian V et al (2023) Risk factors associated with post – SARS-COV-2 condition: a systematic review and meta-analysis. JAMA internal medicine Davis HE et al (2023) Author correction: long COVID: major findings, mechanisms and recommendations. Nat Rev Microbiol 21(6):408–408 Batiha G, El-Saber et al (2022) Pathophysiology of Post-COVID syndromes: a new perspective. Virol J 19(1):158 Eckhardt M et al (2020) A systems approach to infectious disease. Nat Rev Genet 21(6):339–354 Demichev V et al (2021) A time-resolved proteomic and prognostic map of SARS-COV-2. Cell Syst 12(8):780–794 Captur G et al (2022) Plasma proteomic signature predicts who will get persistent symptoms following SARS-CoV-2 infection. EBioMedicine 85 Lu T, Wang Y, Guo T (2022) Multi-omics in SARS-COV-2: Seeing the unseen but overlooked in the clinic. Cell Rep Med 3 .3 Wu M et al (2020) Transcriptional and proteomic insights into the host response in fatal SARS-COV-2 cases. Proceedings of the National Academy of Sciences 117 .45 : 28336–28343 Villar M et al (2021) Characterization by quantitative serum proteomics of immune-related prognostic biomarkers for SARS-COV-2 symptomatology. Front Immunol 12:730710 1Su, Y et al (2022) Multiple early factors anticipate post-acute SARS-COV-2 sequelae. Cell 185 .5 : 881–895 Kroenke, Kurt RL, Spitzer, Janet BW (2001) The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med 16(9):606–613 enkins C, David et al (1988) A scale for the estimation of sleep problems in clinical research. J Clin Epidemiol 41(4):313–321 Nasreddine ZS et al (2005) The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc 53(4):695–699 Zhang J-M (2007) Cytokines, inflammation, and pain. Int Anesthesiol Clin 45(2):27–37 Cavalli G et al (2021) Interleukin 1α: a comprehensive review on the role of IL-1α in the pathogenesis and treatment of autoimmune and inflammatory diseases. Autoimmun rev 20(3):102763 Gregoriou S et al (2010) Cytokines and other mediators in alopecia areata. Mediators of inflammation 2010 Morgun EI, Ekaterina A (2020) Vorotelyak. Epidermal stem cells in hair follicle cycling and skin regeneration: a view from the perspective of inflammation. Front cell Dev biology 8:581697 Van Lenten, Brian J et al (2001) High-density lipoprotein loses its anti-inflammatory properties during acute influenza a infection. Circulation 103 18:2283–2288 Parween F, Rinkoo Devi Gupta (2022) Insights into the role of paraoxonase 2 in human pathophysiology. J Biosci 47(1):4 Maes M et al (2011) Increased plasma peroxides as a marker of oxidative stress in myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS). Med Sci monitor: Int Med J experimental Clin Res 17(4):SC11 Small DM, John Prescott (2005) Odor/taste integration and the perception of flavor. Exp Brain Res 166:345–357 Dumas L-E et al (2024) Impact of post-SARS-COV-2 olfactory disorders on quality of life, hedonic experiences and psychiatric dimensions in general population. BMC Psychiatry 24(1):111 Park H et al (2021) The association between olfactory function and cognitive impairment in older persons with cognitive impairments: a cross-sectional study. Healthcare. Vol. 9 . No. 4. MDPI Lebedeva NN et al (2020) Correlation between Measures of Anxiety and Heart Rate Variability in Men and Women with Affective Disorders. Neurosci Behav Physiol 50:743–749 Steinberger DC, Deanna M (2023) Barch. Investigating the link between depression, cognition, and motivation in late childhood. Child Psychiatry Hum Dev 54(2):570–581 Lv Y et al (2022) Bioinformatics and systems biology approach to identify the pathogenetic link of Long COVID and Myalgic Encephalomyelitis/Chronic Fatigue Syndrome. Front Immunol 13:952987 Jha P, Kumar et al (2021) Gene expression profiling reveals the shared and distinct transcriptional signatures in human lung epithelial cells infected with SARS-CoV-2, MERS-CoV, or SARS-CoV: potential implications in cardiovascular complications of SARS-COV-2. Front Cardiovasc Med 7:623012 Morris G et al (2018) Leaky brain in neurological and psychiatric disorders: Drivers and consequences. Australian New Z J Psychiatry 52(10):924–948 López-Hernández Yamilé et al (2023) The plasma metabolome of long COVID patients two years after infection. Sci Rep 13(1):12420 Bader SM et al (2022) Programmed cell death: the pathways to severe SARS-COV-2? Biochem J 479(5):609–628 Thiemann S, Linda G (2011) Baum. The road less traveled: regulation of leukocyte migration across vascular and lymphatic endothelium by galectins. J Clin Immunol 31:2–9 Louveau A, Mesquita SD, Kipnis J (2016) Lymphatics in neurological disorders: a neuro-lympho-vascular component of multiple sclerosis and Alzheimer’s disease? Neuron 91 .5 : 957–973 Libby P, Lüscher T (2020) COVID-19 is, in the end, an endothelial disease. Eur heart J vol 41(32):3038–3044 Denorme F, Vanhoorelbeke K, Simon F, De Meyer (2019) von Willebrand factor and platelet glycoprotein Ib: a thromboinflammatory axis in stroke. Front Immunol 10:495067 Rubenfire M et al (1986) Platelet hyperaggregability in patients with chest pain and angiographically normal coronary arteries. Am J Cardiol 57:657–660 Allendes FJ et al (2023) Cardiovascular and autonomic dysfunction in long-COVID syndrome and the potential role of non-invasive therapeutic strategies on cardiovascular outcomes. Front Med 9:1095249 Kato A, Schleimer RP (2022) Bleier. Mechanisms and pathogenesis of chronic rhinosinusitis. J Allergy Clin Immunol 149:1491–1503 Parente L, Solito E (2004) Annexin 1: more than an anti-phospholipase protein. Inflamm Res 53:125–132 Younger J, Michael et al (2004) A foldable CFTR∆F508 biogenic intermediate accumulates upon inhibition of the Hsc70–CHIP E3 ubiquitin ligase. J Cell Biol 167(6):1075–1085 Wang Y, Mandelkow E (2016) Tau in physiology and pathology. Nat Rev Neurosci 17(1):22–35 Griffin W, Sue T et al (2006) Interleukin-1 mediates Alzheimer and Lewy body pathologies. J Neuroinflamm 3:1–9 Karlén A et al (2009) Nogo receptor 1 regulates formation of lasting memories. Proceedings of the National Academy of Sciences 106.48 : 20476–20481 Tan BK, Jyn et al (2022) Prognosis and persistence of smell and taste dysfunction in patients with SARS-CoV-2: meta-analysis with parametric cure modelling of recovery curves. bmj 378 Ajmani GS et al (2017) Smoking and olfactory dysfunction: a systematic literature review and meta-analysis. Laryngoscope 127(8):1753–1761 Ding J et al (2023) Association between inflammatory biomarkers and venous thromboembolism: a systematic review and meta-analysis. Thromb J 21(1):82 Alhassan S et al (2017) Clinical presentation and risk factors of venous thromboembolic disease. Crit Care Nurs Q 40(3):201–209 Ogungbe O et al (2022) Cardiac postacute sequelae symptoms of SARS-CoV-2 in community-dwelling adults: cross-sectional study. Open Heart 9(2):e002084 Rahimi K (2020) Self-palpation for detection of paroxysmal atrial fibrillation: Much noise with little signal. PLoS Med 17(3):e1003098 Iadecola C et al (2016) Impact of hypertension on cognitive function: a scientific statement from the American Heart Association. Hypertension 68 .6 : e67-e94 Aung N et al (2020) Causal inference for genetic obesity, cardiometabolic profile and SARS-COV-2 susceptibility: a Mendelian randomization study. Front Genet 11:586308 Baumkötter R et al (2025) Risk tools for predicting long-term sequelae based on symptom profiles after known and undetected SARS-CoV-2 infections in the population. European Journal of Epidemiology , : in revision World Health Organization (‎2021)‎. Global COVID-19 clinical platform case report form (‎CRF)‎ for post COVID condition (‎post COVID-19 CRF)‎, 9 February 2021. World Health Organization Wild P et al (2012) The Gutenberg Health Study. Fed Health Gaz 55:824–830 Arnold N, Koenig W (2021) Polygenic risk score: clinically useful tool for prediction of cardiovascular disease and benefit from lipid-lowering therapy? Cardiovasc Drugs Ther 35(3):627–635 Zhou Y et al (2019) Metascape provides a biologist-oriented resource for the analysis of systems-level datasets. Nat Commun 10(1):1523 Csárdi G, Nepusz T et al (2024) igraph: Network Analysis and Visualization in R Tables Table 1 is available in the Supplementary Files section. Additional Declarations There is NO Competing Interest. Supplementary Files FISUPPProteomicAtlasofPostCOVIDSequelae.docx Proteomic Atlas of Post-COVID Sequelae Table1.png Cite Share Download PDF Status: Under Review 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-6024310","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":417619868,"identity":"d7dc43d1-55e9-4a89-b7fa-c7569cf54af1","order_by":0,"name":"Philipp Wild","email":"data:image/png;base64,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","orcid":"https://orcid.org/0000-0003-4413-9752","institution":"University Medical Center of the Johannes Gutenberg University Mainz","correspondingAuthor":true,"prefix":"","firstName":"Philipp","middleName":"","lastName":"Wild","suffix":""},{"id":417619869,"identity":"2265ee40-ea03-4e6a-afc4-0d2ee1dda2e1","order_by":1,"name":"Sepehr Golriz Khatami","email":"","orcid":"","institution":"Universiätsmedizin Mainz","correspondingAuthor":false,"prefix":"","firstName":"Sepehr","middleName":"Golriz","lastName":"Khatami","suffix":""},{"id":417619870,"identity":"d48923bb-ec8d-4034-8702-e9e0213d3b1a","order_by":2,"name":"Rieke Baumkötter","email":"","orcid":"","institution":"Universiätsmedizin Mainz","correspondingAuthor":false,"prefix":"","firstName":"Rieke","middleName":"","lastName":"Baumkötter","suffix":""},{"id":417619871,"identity":"c10d2fd1-6714-481c-99a4-81d2fed8e416","order_by":3,"name":"Thomas Koeck","email":"","orcid":"","institution":"University Medical Center of the Johannes Gutenberg University Mainz","correspondingAuthor":false,"prefix":"","firstName":"Thomas","middleName":"","lastName":"Koeck","suffix":""},{"id":417619872,"identity":"dce5524f-b721-4c29-8cff-72684fd98700","order_by":4,"name":"Ayesha Syed Mamoor Alam","email":"","orcid":"","institution":"Universiätsmedizin Mainz","correspondingAuthor":false,"prefix":"","firstName":"Ayesha","middleName":"Syed Mamoor","lastName":"Alam","suffix":""},{"id":417619873,"identity":"9b7926c6-7e22-44cf-9974-afc372264e9f","order_by":5,"name":"Simge Yilmaz","email":"","orcid":"","institution":"Universiätsmedizin Mainz","correspondingAuthor":false,"prefix":"","firstName":"Simge","middleName":"","lastName":"Yilmaz","suffix":""},{"id":417619874,"identity":"445ed6d6-8d55-4a45-9ec9-ee1a8b9201e8","order_by":6,"name":"Julian Chalabi","email":"","orcid":"","institution":"Universiätsmedizin Mainz","correspondingAuthor":false,"prefix":"","firstName":"Julian","middleName":"","lastName":"Chalabi","suffix":""},{"id":417619875,"identity":"9a955814-19de-4bfb-903c-678c3e672843","order_by":7,"name":"Steffen Rapp","email":"","orcid":"","institution":"University Medical Center of the Johannes Gutenberg University Mainz","correspondingAuthor":false,"prefix":"","firstName":"Steffen","middleName":"","lastName":"Rapp","suffix":""},{"id":417619876,"identity":"578fb8f0-359b-4d3b-bcc0-bf725d738352","order_by":8,"name":"Federico Marini","email":"","orcid":"https://orcid.org/0000-0003-3252-7758","institution":"University Medical Center of the Johannes Gutenberg-University Mainz","correspondingAuthor":false,"prefix":"","firstName":"Federico","middleName":"","lastName":"Marini","suffix":""},{"id":417619877,"identity":"dba38fc8-d82c-440a-9e0c-a315ac4e20b9","order_by":9,"name":"Markus Schepers","email":"","orcid":"https://orcid.org/0000-0003-2436-1540","institution":"Universiätsmedizin Mainz","correspondingAuthor":false,"prefix":"","firstName":"Markus","middleName":"","lastName":"Schepers","suffix":""},{"id":417619878,"identity":"3ed2ed55-8faf-48ed-9b98-db9fa534974d","order_by":10,"name":"Tobias Bopp","email":"","orcid":"https://orcid.org/0000-0002-3232-8065","institution":"University Medical Center of the Johannes Gutenberg University Mainz","correspondingAuthor":false,"prefix":"","firstName":"Tobias","middleName":"","lastName":"Bopp","suffix":""},{"id":417619879,"identity":"846c56e8-6920-4d84-a054-d18e82ffaded","order_by":11,"name":"Stefan Bittner","email":"","orcid":"","institution":"Universiätsmedizin Mainz","correspondingAuthor":false,"prefix":"","firstName":"Stefan","middleName":"","lastName":"Bittner","suffix":""},{"id":417619880,"identity":"41d701e2-3db0-4d63-bcee-c008a9a91153","order_by":12,"name":"Alexander Schuster","email":"","orcid":"","institution":"University Medical Center Mainz","correspondingAuthor":false,"prefix":"","firstName":"Alexander","middleName":"","lastName":"Schuster","suffix":""},{"id":417619881,"identity":"2ef15304-e741-4433-93c2-cf3a15947bd9","order_by":13,"name":"Manfred E. Beutel","email":"","orcid":"","institution":"Universiätsmedizin Mainz","correspondingAuthor":false,"prefix":"","firstName":"Manfred","middleName":"E.","lastName":"Beutel","suffix":""},{"id":417619882,"identity":"d142a51c-4185-4af1-9c6e-42f60efc349f","order_by":14,"name":"Thomas Munzel","email":"","orcid":"https://orcid.org/0000-0001-5503-4150","institution":"Johannes Gutenberg University Mainz","correspondingAuthor":false,"prefix":"","firstName":"Thomas","middleName":"","lastName":"Munzel","suffix":""},{"id":417619883,"identity":"83d66fe0-1725-4d83-8915-6c87301819a0","order_by":15,"name":"Karl Lackner","email":"","orcid":"","institution":"University Medical Center Mainz","correspondingAuthor":false,"prefix":"","firstName":"Karl","middleName":"","lastName":"Lackner","suffix":""},{"id":417619884,"identity":"97084b3c-02c5-4258-9e5c-60a0c5f5c953","order_by":16,"name":"Vincent Ten Cate","email":"","orcid":"","institution":"University Medical Center of the Johannes Gutenberg University Mainz","correspondingAuthor":false,"prefix":"","firstName":"Vincent","middleName":"Ten","lastName":"Cate","suffix":""}],"badges":[],"createdAt":"2025-02-13 15:17:37","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6024310/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6024310/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":77769213,"identity":"c28375fc-9476-4d15-90c9-0a926dcfa363","added_by":"auto","created_at":"2025-03-05 10:32:04","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":251057,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe predictive power of machine learning models.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe two right columns show the number of selected proteins and the prevalence of the symptoms in the investigated cohort, respectively. The light blue bar shows the differentiation ability of the model in forecasting post-COVID, while the dark blue ones show the symptom-specific model’s classification performance. The six lowest-ranked symptoms exhibited limited discriminative ability for differentiating individuals with and without post-COVID symptoms.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6024310/v1/f06d6d781abdb65ace685a28.png"},{"id":77767286,"identity":"ab2115c1-4c60-4ede-a1e0-28548fb13a8a","added_by":"auto","created_at":"2025-03-05 10:16:03","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":877931,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eProteins-associated symptoms and protein overlap between symptoms.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe 15 symptoms and their associated proteins are shown in different colors around the circumference of the oval. The small circles and their size in front of the proteins show that specific proteins are common between at least three symptoms. The smaller the circles, they are associated with less number of symptoms. The enclosed numbers represent the frequency of protein sharing among different symptoms. The curved lines connect the common proteins across symptoms.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-6024310/v1/1e4d9e6eefac3384dc0a87a4.png"},{"id":77767287,"identity":"a2ea633d-0e98-4b23-8373-faa55b8b5149","added_by":"auto","created_at":"2025-03-05 10:16:04","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":520310,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRelation between post-COVID symptoms.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA) \u003c/strong\u003eThe correlation between protein scores of 15 different symptoms\u003cstrong\u003e. B) \u003c/strong\u003eAssociation between the symptoms' correlation and their coexistence. \u003cstrong\u003eC) \u003c/strong\u003ePotential mechanisms underpinning post-COVID symptoms. The circles with different dark colors are shown as clusters composing different symptoms grouped based on the correlations summarized in part A. The circles with distinct light colors represent the pathways associated with specific clusters.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-6024310/v1/2dc43a5dd769d6320c27a121.png"},{"id":77767285,"identity":"46254345-9f18-42af-a5c5-453199528cb9","added_by":"auto","created_at":"2025-03-05 10:16:03","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":475411,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAssociations between symptoms-related proteins and objective clinical tests.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe x-axis shows the level of associations and the y-axis depicts the different symptoms. The associations were grouped based on distinctive clinical tests. The color of the bars represents different significance levels (alpha thresholds) used for statistical analysis.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-6024310/v1/9302299196fda8932288441b.png"},{"id":77768566,"identity":"79e3ab9c-a648-4848-b768-1956502ab3d7","added_by":"auto","created_at":"2025-03-05 10:24:04","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":409200,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe relevance of disease risk factors and pre-existing conditions to post-COVID symptoms.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe statistically significant associations between the symptoms (x-axis) and comorbidities as well as risk factors (y-axis) are highlighted in different colors based on the direction of association; Positive associations: a range of blue colors, negative associations: a range of orange colors. The asterisks (*) depict various alpha-level thresholds.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-6024310/v1/676db464ba864c0d6516fb50.png"},{"id":77767291,"identity":"aaa25f7f-a50f-4025-ad98-a388833feb11","added_by":"auto","created_at":"2025-03-05 10:16:04","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":319407,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eInterrelation between genetic predisposition and post-COVID symtoms development.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePRSs have been used as proxies for analysis of the association between 10 different genetic predispositions (y-axis) and 15 post-COVID symptoms (x-axis). The statistically significant associations are highlighted with light green circles and the number of * in the circles shows different thresholds for alpha-level.\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-6024310/v1/58ed978d5bb2fc96a631fcc1.png"},{"id":77770213,"identity":"17d471ed-4ec2-45b9-aad6-7b110bd08f56","added_by":"auto","created_at":"2025-03-05 10:40:06","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3734324,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6024310/v1/81f1410c-9c9d-4ecd-8aa8-3ab4f45ad430.pdf"},{"id":77767289,"identity":"4c6f4a25-2161-4fb9-93f6-f7bbe0cdb868","added_by":"auto","created_at":"2025-03-05 10:16:04","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":24929,"visible":true,"origin":"","legend":"Proteomic Atlas of Post-COVID Sequelae","description":"","filename":"FISUPPProteomicAtlasofPostCOVIDSequelae.docx","url":"https://assets-eu.researchsquare.com/files/rs-6024310/v1/6cdd8203b0dc75bfce063acd.docx"},{"id":77767284,"identity":"f134096c-360b-48f2-8976-85de2d453fcb","added_by":"auto","created_at":"2025-03-05 10:16:03","extension":"png","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":96025,"visible":true,"origin":"","legend":"","description":"","filename":"Table1.png","url":"https://assets-eu.researchsquare.com/files/rs-6024310/v1/9545cf39a2a5ce26b11e621b.png"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Proteomic Atlas of Post-COVID Sequelae","fulltext":[{"header":"INTRODUCTION","content":"\u003cp\u003eAmid the ongoing global SARS-COV-2 infections, the emergence of post-COVID sequelae (PCS), affecting 30\u0026ndash;50% of COVID-19 survivors, presents novel challenges.[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e] PCS is a syndrome characterized by a wide range of symptoms, from more common ones, including fatigue, to more severe manifestations such as neurological disorders.[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e] PCS is defined by the World Health Organization (WHO) as the emergence of new symptoms or the persistence of symptoms beyond three months following a SARS-COV-2 infection in the absence of an alternative explanation.[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e] Various factors such as demographic attributes (e.g., age and sex), pre-existing somatic and cognitive conditions, and history of hospitalization for SARS-COV-2, including admission to an intensive care unit, are linked with an increased risk of PCS.[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e] Several putative pathophysiologic mechanisms, namely viral persistence, hypercoagulopathy, immune dysregulation, autoimmunity, hyperinflammation, imbalance of the renin-angiotensin system, and oxidative stress, have been proposed as causing PCS.[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]\u003c/p\u003e \u003cp\u003eThe advent of high-throughput proteomic assays enabled the measurement of thousands of proteins, facilitating the systematic examination of molecular changes throughout the disease and providing comprehensive insight into the biological mechanisms and disease etiology.[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e] Proteomics has been widely employed in SARS-COV-2 research.[\u003cspan additionalcitationids=\"CR10 CR11 CR12\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e] For instance, an increase in protein expression level related to neutrophils or the complement system was found in severe disease courses.[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e] Proteins regulating the circadian rhythm pathway were identified in higher concentrations in post-COVID patients with neurological abnormalities [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Proteomics was also employed to develop a classification model to predict long COVID using proteomic data from acute-phase SARS-COV-2 patients with 12-month symptom persistence.[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e] Although these studies shed light on some aspects of the syndrome, their reach and impact are limited as brief follow-up periods, small cohort sizes, suboptimal protein coverage, and a lack of comprehensive long-term data hindered the ability to draw definitive conclusions.\u003c/p\u003e \u003cp\u003eThis study leveraged a high-throughput proteomic technology by proximity extension assay technology (Olink, Uppsala, Sweden) to generate comprehensive protein profiles of SARS-COV-2 survivors from plasma samples of a population-based cohort. The aim was to employ machine-learning models to identify protein signatures associated with developing of post-COVID symptoms. Symptoms-associated proteins were investigated concerning objective clinical tests, encompassing cognitive, psychiatric, and somatic assessments, to explore the potential clinical utility of the detected proteins. The work also capitalized on the synergy between multi-automated data and comprehensive phenotypic information to characterize key biological factors associated with PCS heterogeneity.\u003c/p\u003e"},{"header":"RESULTS","content":"\u003ch2\u003eSample characteristics\u003c/h2\u003e\n\u003cp\u003eThe clinical characteristics of 495 SARS-COV-2 survivors from GCS included in this study, with a mean age of 53.2\u0026plusmn;15.8 years and an approximately balanced sex ratio (47.9% female) are shown in \u003cstrong\u003eTable 1\u003c/strong\u003e. Among cardiovascular risk factors, arterial hypertension emerged as the most prevalent, affecting 43.8% of the study sample whereas type 2 diabetes mellitus and smoking were observed with lower frequencies. Similarly, pre-existing cardiovascular disease was the most common comorbidity, with a prevalence of 12.7%. Chronic liver disease and atrial fibrillation (AF) were less frequent, affecting 2.0% and 2.2% of participants, respectively.\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eSymptoms of post-COVID and machine learning-derived protein signatures\u003c/h2\u003e\n\u003cp\u003eStudy participants were asked about 61 distinct symptoms (see \u003cstrong\u003eSupplementary Methods\u003c/strong\u003e). The symptom catalogue was used for the development of protein-based models explaining the presence of post-COVID or specifically symptoms. All symptoms reported by fewer than 10 individuals were excluded from subsequent analysis to ensure the reliability and robustness of predictive models. Overall this led to 21 symptoms to be further explored. Fatigue was the most prevalent symptom, affecting 11.7% of individuals (N=58), while slow movements (2.0%; n=10) and depression (2.2%; N=11) were less frequent manifestations within the sample.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eUpon completion of the quality control, 46 proteins from an initial set of 2,945 were excluded due to being below the limit of detection, low call rates, or technical variability. Of 2,899 remaining proteins, 235 unique ones were associated with 15 distinct post-COVID symptoms.\u0026nbsp;The model for predicting the presence of post-COVID regardless of the symptoms present, exhibited a lower discriminative ability (leave-one-out cross validation; LOOCV AUC = 0.62) compared to the symptom-specific predictive models. Among the models, each tailored to a specific symptom, the one for predicting chest pain as a long-term sequela of SARS-CoV-2 exhibited the highest LOOCV AUC of 0.84, while the one capturing anxiety showed the lowest discriminative ability with an AUC of 0.51. The predictive models could not differentiate between individuals who exhibited the following symptoms and those who did not, across six distinct symptoms namely balance issues, cough, joint pain, and muscle pain, headache, and sleep disturbances (\u003cstrong\u003eFigure 1\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eThe \u0026lsquo;shortness of breath\u0026rsquo; symptom had the highest number of associated proteins, totaling N=51. Conversely, the symptoms \u0026lsquo;concentration problems\u0026rsquo; with counts of five proteins including\u0026nbsp;protein kinase AMP-activated non-catalytic subunit gamma 3 (PRKAG3), Ketohexokinase (KHK), Calcineurin like EF-hand protein 1 (CHP1), Cell division cycle 25A (CDC25A), Annexin A1 (ANXA1) and \u0026lsquo;anxiety\u0026rsquo; with three (Interleukin 1 alpha; IL1A, ANXA1, Switching B cell complex subunit; SWAP70) had the fewest proteins. Interestingly, some proteins were associated with several symptoms, the connection between which also showed clinical plausibility (\u003cstrong\u003eFigure 2\u003c/strong\u003e): IL1A is the most frequently detected protein, associated with five distinct symptoms, including shortness of breath, hair loss, depression, anxiety, and memory problems. Disturbances in olfaction and taste were found to have the highest number of shared proteins (N=11), with Toll like receptor 4 (TLR4), Thyroglobulin (TG), Melanotransferrin (MELFT), Creatine kinase, mitochondrial 1A (CKMT1A), Chymotrypsin like (CTRL), Eukaryotic translation initiation factor 1A X-linked (EIF1AX), Apolipoprotein B receptor (APOBR), Sphingomyelin phosphodiesterase acid like 3B (SMPDL3B), Asporin (ASPN), and TSPY like 1 (TSPYl1), followed by shortness of breath and fatigue (CD3 gamma subunit of T-cell receptor complex; CD3G, Pancreatic lipase related protein 2; PNLIPRP2, Sulfatase modifying factor 2; SUMF2, Endosome associated trafficking regulator 1; ENTR1) as well as shortness of breath and depression (MANSC domain containing 4; MANSC4, IL1A,\u0026nbsp;S100 calcium binding protein A11; S100A11, and\u0026nbsp;Paraoxonase 3; PON3) each sharing four proteins.\u003c/p\u003e\n\u003cp\u003eTo elucidate the relationship between the symptom-specific protein signatures a correlation matrix was constructed (\u003cstrong\u003eFigure 3A\u003c/strong\u003e): The strongest positive correlations were observed for slow movements and loss of interest (r=0.82) and for anxiety and palpitations (r=0.75). In contrast, the most pronounced inverse correlations were identified for mood swings and olfactory disturbances (r=-0.21) as well as for concentration problems and olfactory disturbances (r=-0.19). Furthermore, the analysis revealed higher positive correlations between fatigue and slow movements (r=0.57), concentration and memory problem (r=0.53) and olfactory and taste disturbances (r=0.52). Subsequently, the co-prevalence of clinical post-COVID symptoms was compared with the correlation of the respective symptom-related protein scores and showed a moderate positive correlation of 0.36 (\u003cstrong\u003eFigure 3B\u003c/strong\u003e).\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003ePutative pathomechanisms underlying post-COVID related symptoms\u003c/h2\u003e\n\u003cp\u003eTo gain a deeper understanding of the biological mechanisms underlying post-COVID symptoms, first, hierarchical clustering was employed to group symptoms based on the correlation of their corresponding protein signatures. Proteins associated with the co-clustered symptoms were then subjected to pathway enrichment analysis. The clustering analysis yielded six discrete symptom clusters, as illustrated in Figure 3C: cluster 1 with fatigue/tiredness, shortness of breath, loss of interest or pleasure, and slow movements; cluster 2 with anxiety, and palpitations; cluster 3 with hair loss, memory problems, mood swings, and concentration problems; cluster 4 with olfactory disturbances, and taste disturbances; cluster 5 with sleep disturbances, and depression; and cluster 6 with chest pain.\u003c/p\u003e\n\u003cp\u003eA total of\u0026nbsp;59 distinct molecular pathways were identified as relevantly associated with symptom clusters. The largest number of molecular pathways were identified in symptom cluster 1 (n = 20), while the smallest number were identified in symptom cluster 6 (n = 1). The most prevalent class of pathways within symptom cluster 1 were those pertaining to metabolism, including fatty acid metabolism and the regulation of lipid transport. Additionally, pathways associated with signalling, such as those involving ATM signalling and P53 downstream, were also highly represented. The most common pathways associated with symptom cluster 2 were those related to cellular stress and damage response, including pathways involved in the response to growth factors and temperature stimuli.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn contrast, metabolic pathways were the most frequent category of pathways associated with symptom cluster 4. The most frequently occurring pathways associated with symptom cluster 3 were those pertaining to the immune and inflammatory responses and those involved in signalling. The sole pathway found to be associated with symptom cluster 6 is a cell proliferation and regulation pathway (i.e., positive regulation of exocytosis).\u0026nbsp;While the enrichment analysis revealed distinct pathways associated with individual clusters, several pathways were identified as shared across multiple symptom clusters, notably the \u0026quot;response to temperature stimulus\u0026quot; (symptom clusters 1, 2, 3), \u0026quot;carbohydrate derivative catabolic process\u0026quot; (symptom clusters 1, 4), and \u0026quot;response to wounding\u0026quot; (symptom clusters 2, 3).\u003c/p\u003e\n\u003ch2\u003eClinical validation of post-COVID symptom-specific protein signatures \u003c/h2\u003e\n\u003cp\u003eTo further explore the potential clinical utility of proteins discovered for individual post-COVID symptoms, their link with measurements from various established and objective clinical tests, encompassing cognitive, psychiatric, and somatic assessments were analyzed (\u003cstrong\u003eFigure 4\u003c/strong\u003e). Broadly, a correlative relationship between symptom-centered objective clinical tests and the protein score associated with the corresponding symptom was detected. For example, a positive association was found between protein signature related with mood swings and the nine-item Patient Health Questionnaire (PHQ-9)\u0026nbsp;\u003cstrong\u003e[15]\u003c/strong\u003e, a depressive symptom scale. Similarly, the sleep disturbances- protein signature was associated with the Jenkins Sleep Scale (JSS)\u003cstrong\u003e\u0026nbsp;[16]\u003c/strong\u003e, a scale for estimating of sleep problems in clinical research. In contrast, a negative association was identified between the Montreal Cognitive Assessment (MoCA)\u0026nbsp;\u003cstrong\u003e[17]\u003c/strong\u003e as a screening measure for the diagnosis of cognitive impairment and the protein signatures corresponding to cognitive function, specifically memory with regression coefficients (\u0026beta; = -0.11) and concentration with (\u0026beta; \u0026asymp; -0.1). Similarly, an analogous association was detected between the objective clinical olfactory test (i.e., a smell test comprising 12 smelling sticks) and the protein signatures indicative of olfactory and gustatory modalities, individually. However, no clinically relevant link was found between spirometry tests measuring vital capacity (VC) and forced expiratory volume in one second (FEV1) and the protein signature for shortness of breath.\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eRelation of the clinical patient profile with post-COVID symptom-specific protein signatures \u003c/h2\u003e\n\u003cp\u003eThe association between pre-existing conditions and protein signatures of individual symptoms was investigated to characterize the clinical factors, like risk factors and comorbidities, that may be predisposing to or even be involved in the development of post-COVID symptoms, the association between pre-existing conditions and the protein signatures of individual symptoms was investigated (\u003cstrong\u003eFigure 5\u003c/strong\u003e). The analyses revealed the strongest positive association between diabetes mellitus (DM) and the protein signature associated with memory problems (\u0026beta;\u003csub\u003eper SD\u003c/sub\u003e = 0.28, 90%Cl [-0.05; 0.6]), while the strongest inverse association was detected between obesity and the protein signature for hair loss (\u0026beta;\u003csub\u003eper SD\u003c/sub\u003e = -0.28, 95%Cl [-0.54; -0.03]). Positive associations were observed between smoking and protein signatures linked to taste (\u0026beta;\u003csub\u003eper SD\u003c/sub\u003e = 0.21, 90%Cl [0; 0.41]) and olfactory disturbances (\u0026beta;\u003csub\u003eper SD\u003c/sub\u003e = 0.16, 90%Cl [-0.02; 0.35]). Furthermore, the study highlighted the influence of comorbidities on the prevalence of diverse post-COVID symptoms, as indicated by the protein signatures. The protein signature of sleep disturbances exhibited an inverse association with the largest number of comorbidities including coronary artery disease (CAD), peripheral arterial disease, history of pulmonary embolism, congestive heart failure (CHF), and arterial fibrillation (AF). The strongest association was observed for CHF (\u0026beta; = 1.38, 95%Cl 0.34; 2.43]). In contrast, the protein signatures of the post-COVID symptoms namely anxiety, memory problems, palpitations, olfactory disturbances, chest pain, loss of interest and slow movements were only associated with one comorbidity each. These were a history of venous thromboembolism (VTE), CAD, AF, history of stroke and chronic obstructive pulmonary disease (COPD). Furthermore, the investigation demonstrated that the strongest positive standardized association was observed for the relation of H.x VTE with the protein signature of shortness of breath (\u0026beta;\u003csub\u003eper SD\u003c/sub\u003e=1.48, 95%Cl [-1.82; 0.64]), while the strongest inverse association was noted for COPD with the signature of loss of interest (\u0026beta;\u003csub\u003eper SD\u003c/sub\u003e=-1.3, 90%Cl [-0.35; 1.46]).\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eGenetic predisposition for cardiovascular risk factors and diseases and susceptibility of post-COVID symptoms\u003c/h2\u003e\n\u003cp\u003eFinally, the potential impact of a genetic predisposition for cardiovascular risk factors and diseases on the susceptibility to post-COVID symptoms was evaluated using polygenic risk scores and the protein signatures of post-COVID symptoms\u0026nbsp;(\u003cstrong\u003eFigure 6\u003c/strong\u003e). Several associations were identified between polygenic risk scores and various post-COVID manifestations: The most substantial positive association was observed between the polygenic risk scores for cholesterol concentration and the \u0026quot;loss of interest\u0026quot; proteins (\u0026beta; = 0.36). Conversely, the highest negative associations were found between the hyperlipidaemia-computed polygenic risk score and the protein signatures for \u0026quot;slow movement\u0026quot; (\u0026beta; = -0.23) and for \u0026ldquo;memory problem\u0026quot; (\u0026beta; = -0.21). A positive correlation was detected between the polygenic risk score for stroke and the protein signature for chest pain (\u0026beta; = 0.25) as well as between the cholesterol-estimated polygenic risk score and the proteins linked to fatigue (\u0026beta; = 0.20). Conversely, an inverse association was observed between the protein signature for olfactory disturbance and the polygenic risk score for smoking (\u0026beta; = -0.1).\u003c/p\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eThis study generated protein signatures of the 15 most common symptoms of the post-COVID, investigated their interrelations, and explored the putative underlying pathomechanisms. The analysis of associations between the symptom-specific protein signatures and objective clinical tests demonstrated the potential clinical utility of detected proteins. The study also examined the relevance of pre-existing conditions and disease risk factors with symptom-related protein scores, and the link between genetic predisposition of clinical traits and disease susceptibility with protein scores of individual symptoms. Overall, machine learning models demonstrated superior discriminative capabilities concerning individual symptoms, as opposed to the protein profile associated with the general presence of a post-COVID. A more comprehensive understanding may therefore be achieved by deconstructing the syndrome into distinct symptom clusters, each potentially linked to different molecular mechanisms.\u003c/p\u003e \u003cp\u003eConcerning protein-symptom associations, Interleukin-1 alpha (IL1A) emerged as the most frequently identified protein. It was associated with five distinct clinical symptoms: dyspnea, hair loss, depression, anxiety, and cognitive impairments. IL1A, a known proinflammatory mediator, crucial role modulating inflammation and SARS-COV-2 pathogenesis.[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e] Studies have shown that SARS-COV-2 infection increases interleukin-1 (IL-1) expression leading to inflammatory factor accumulation in the lungs and potentially triggering a cytokine storm and chronic inflammation.[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e] This persistent inflammatory condition may be associated with immunothrombosis and the formation of microclots, impede oxygen exchange, and contribute to shortness of breath. Similarly, inflammatory cytokines such as IL-1 can disrupt the hair growth cycle, causing increased shedding and hair loss [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e] and hinder hair regrowth.[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e] Additionally, Paraoxonase 3 (PON3) was also associated with a cluster of symptoms, including fatigue, slow movements, depression, mood swings, and loss of interest. The extensive inflammation observed during SARS-COV-2 infection dysregulates the expression of the paraoxonase family (PON1, PON2, PON3)[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. This dysregulation can lead to increased oxidative stress, as PON enzymes play a crucial role in protecting against it by detoxifying lipid peroxides and preventing low-density lipoprotein (LDL) oxidatio [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e], in which elevated LDL cholesterol has been associated with symptoms such as fatigue, depression, and diminished interest. [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]\u003c/p\u003e \u003cp\u003eSeveral symptoms were co-prevalent, including the well-known co-prevalence of olfactory and taste disturbances. Various studies reported an intimate functioning between the chemosensory systems, specifically smell and taste, and about 30% of SARS-COV-2-infected individuals experienced a dysfunction of both systems.[\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e] Contrary to previous studies [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e], but in concordance with Park \u003cem\u003eet al.\u003c/em\u003e [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e], an inverse relationship between olfactory function and concentration problems was observed. The correlation identified between mental and somatic symptoms, particularly anxiety and palpitations, is supported by studies showing that patients with affective disorders often have symptoms of autonomic vascular dystonia, including severe palpitations. [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e] This also relates to the link between concentration problems, memory problems, and mood, as depression often leads to difficulties in remembering information, especially positive information [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe pathway enrichment analysis revealed a wide range of pathway categories from metabolic and signaling pathways to cellular stress and damage response pathways, as well as immune and inflammatory pathways. Among pathways associated with symptoms in cluster 1, the ATM signaling pathway, signaling by interleukins, fatty acid metabolism, and apoptotic signaling pathway were found, all of which play roles in the development fatigue/tiredness [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e], shortness of breath [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e], loss of interest and pleasure [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e], or slow movements [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. Different mechanisms have been proposed for how apoptotic signaling pathway contributes to the development and manifestation of post-COVID, namely persistent inflammation and tissue damage, dysregulated immune responses, autoimmunity, and mitochondrial dysfunction. [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e] The regulation of leukocyte migration and the regulation of kinase activity are two of the pathways linked to symptoms in cluster 3. Studies have shown that SARS-COV-2 is associated with abnormal leukocyte migration.[\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e] Recently lymphatic vessels in the membranes around the brain were discovered and the lymphatic system was introduced as a potential new player in complex neurological problems such as memory impairment.[\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e] The analysis of cluster 6 resulted in the regulation of exocytosis as a pathway associated with the sole symptom chest pain in this cluster. The known endothelial cell inflammation (endotheliitis) in SARS-COV-2 [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e] was reported to lead to degranulation and exocytosis of Weibel-Palade bodies containing von Willebrand factor, promoting recruitment and aggregation of platelets[\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e], which in turn, may increase the risk of thrombosis and subsequent chest pain, specifically in the case of heart or lung affection. [\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]\u003c/p\u003e \u003cp\u003eThe co-existence of diverse symptoms within a single cluster may hint at shared pathophysiological mechanisms across these symptoms. This is supported by the pathway enrichment analysis, which revealed common pathways enriched within each cluster and pathways shared across distinct clusters. For instance, the response to temperature stimulus of clusters 1, 2, and 3 explains the dysregulation of temperature response pathways, mediated through neuroinflammation, cytokine signaling, and autonomic dysfunction. [\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]\u003c/p\u003e \u003cp\u003eThe association between protein-specific molecular signatures and quantitative clinical assessments underscores their diagnostic and therapeutic potential. Several proteins have been identified as being associated with olfactory dysfunction in post-COVID-19 symptoms, including Toll-Like Receptor 4 (TLR4), a mediator of inflammatory responses within the olfactory epithelium; Annexin A1 (ANXA1), an anti-inflammatory mediator; Surfactant Protein A1 (SFTPA1), a component of mucosal immunity; Ring Finger Protein 5 (RNF5), a regulator of protein quality control; Prostaglandin E Synthase 2 (PTGES2), involved in inflammatory pathways; and Nudix Hydrolase 15 (NUDT15), a metabolic regulator. These proteins have been investigated as potential therapeutic targets for olfactory disorders. [\u003cspan additionalcitationids=\"CR42\" citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e] Similarly, the analysis identified proteins associated with memory impairment, including Microtubule-Associated Protein Tau (MAPT), a critical contributor to neurodegenerative pathologies; Interleukin 1 Alpha (IL1A), a mediator of neuroinflammation; Neurotrophin-3 (NTF3), essential for synaptic plasticity; Sex Hormone-Binding Globulin (SHBG), which influences cognitive function; Allograft Inflammatory Factor 1 (AIF1), associated with neuroinflammatory responses; and Reticulon 4 Receptor (RTN4R), a regulator of synaptic remodeling. These proteins represent promising therapeutic targets for the treatment of memory dysfunction, as supported by prior research. [\u003cspan additionalcitationids=\"CR45\" citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e]\u003c/p\u003e \u003cp\u003eThis study also displayed associations between risk factors that may be involved in developing post-COVID symptoms, such as smoking and taste and smell disturbances. Approximately 5% of individuals who report initial chemosensory dysfunction, specifically taste and smell, may experience persistent symptoms six to twelve months following the infection.[\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e] The impact of SARS-COV-2 on the sensory functions, could be intensified by the known influence of smoking on the olfactory and gustatory senses [\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e]. Further interesting links were observed between comorbidities and post-COVID symptoms, including VTE and chest pain, but also AF and palpitation. VTE-induced inflammation with VTE as both an initiator and a perpetrator of inflammation [\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e], may contribute to acute chest discomfort, potentially arising from pleurisy [\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e]. Individuals with AF and hypertension exhibited a two-fold increased risk of developing long-term cardiac complications following a SARS-COV-2 infection [\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e], potentially induced by inflammatory response, dysfunction of the autonomic nervous system, and electrolyte disturbances. These could trigger electrical and structural remodeling of the atria, thereby intensifying SARS-COV-2-induced palpitations, particularly in individuals with AF. [\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e]\u003c/p\u003e \u003cp\u003eInteresting associations were detected between genetics by polygenetic risk scores and post-COVID sequelae: Notably, a genetic predisposition for arterial hypertension was associated with the protein signature indicative of concentration problems. While the precise mechanisms between hypertension and cognitive impairment remain elusive, it is postulated that hypertension could precipitate cognitive decline through alterations in blood vessels or brain damage related to vascular complications, inflammation, and oxidative stress.[\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e] The link between the genetic susceptibility for higher cholesterol concentrations and the protein pattern for fatigue is supported by earlier reports that suggest that a genetic predisposition to hypercholesterolemia may have an increased vulnerability to post-COVID symptome, particularly fatigue.[\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e]\u003c/p\u003e \u003cp\u003eThe current study has several strengths, including the larger, well-characterized cohort with a longer observation period including a pre-pandemic phase, the measurements with proximity extension technology with over 3000 proteins (Olink explore, Uppsala, Sweden) per individual, and the availability of genetic data. However, some relevant limitations must also be taken into account: First, the work focused on quantifying protein concentrations without considering the conformational state of the proteins. This encompasses their folding patterns and three-dimensional structures and is thus relevant for their functionality and potential interactions within the biological system. Second, the study did not analyze protein dynamics, which limits the conclusions that can be drawn. Third, no replication could be implemented in this exploratory study due to the unique setting of this biobank. Fourth, this study was conducted during the dominance of the wild-type and Alpha variants of SARS-CoV-2 and the results may only be valid for these specific variants. Fifth, while the study encompasses several hundred participants, the subset of individuals reporting a specific post-COVID symptom is still small, limiting the power to identify robust predictors or specific classifications. Finally, mechanistic findings, such as chronic inflammation, may be attributed to underlying pathologies or comorbidities that are not directly relevant to post-COVID. The exclusion of these confounding variables is again limited by the subgroups' relatively small size.\u003c/p\u003e \u003cp\u003eThis comprehensive study indicates that PCS represents a condition with diverse and yet unexplored underlying pathophysiological mechanisms. The findings shed light on the complex nature of the syndrome, highlighting that factors such as genetic predisposition, medical history, and disease risk factors play a significant role in the observed heterogeneity of the condition. The protein patterns indicate a more comprehensive understanding could be achieved by deconstructing the syndrome into distinct symptom clusters, each linked to different molecular mechanisms. This approach could potentially facilitate the development of targeted therapies and personalized treatment strategies.\u003c/p\u003e"},{"header":"MATERIALS AND METHODS","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\n \u003ch2\u003eStudy design and population\u003c/h2\u003e\n \u003cp\u003eThis study mainly leveraged post- COVID data from two large-scale, prospective, population-based cohorts, namely the Gutenberg COVID-19 Study (GCS) and the Gutenberg Post COVID Study (GPC) and background information from Gutenberg Health Study (GHS) [\u003cspan class=\"CitationRef\"\u003e15\u003c/span\u003e]. The individual cohorts provide complementary information that was fused to provide more detailed information on the subjects. This analysis employed i) details on post-infection protein profiles and post-COVID symptoms contributed by GCS, ii) deep (sub)clinical phenotypes measured by objective clinical tests in the course of medical-technical examinations offered by GPC for individuals who have contracted SARS-COV-2 infection, and iii) the wealth of baseline information (i.e., comorbidities and risk factors before the pandemic and genetic data provided by GHS. A detailed description of ethics approval and consent to participate is available in the supplementary information.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\n \u003ch2\u003eGutenberg-COVID19 Study\u003c/h2\u003e\n \u003cp\u003eGCS was initiated in 2020 to explore the epidemiological dynamics and population-level impacts of the pandemic across both infected and uninfected population. The GCS enrolled 10,250 participants, 8,121 from GHS aged 45\u0026ndash;88 and 2,129 newly recruited from a random younger sample aged 25\u0026ndash;44. The cohort maintained a balanced sex distribution, with 50.8% of participants being female. Comprehensive biospecimen collection, including various blood fractions (serum, EDTA plasma, PBMC), stool, tooth pocket swabs, and tear fluid samples and routine laboratory parameters and detailed clinical information (including medical history and family history) were obtained from GCS participants at two time points, approximately four months apart. SARS-CoV-2 infection status was determined using a combination of diagnostic methods: i\u003cem\u003e)\u003c/em\u003e quantitative reverse transcription polymerase chain reaction (RT-qPCR), \u003cem\u003eii)\u003c/em\u003e serological testing for anti-SARS-COV-2 antibodies using two immunoassays covering different antibody spectra i.e., the Abbott Architect SARS-COV-2 IgG (Germany) assay and the Roche Elecsys\u0026reg; Anti-SARS-COV-2 (Germany) assay, and iii) self-reported SARS-CoV-2 infection based on computer-assisted personal interview and weekly smartphone-based assessments in the follow-up period. Subsequently, individuals who were SARS-COV-2-positive and a control group without infection that was matched for demographic characteristics (i.e., age and sex) [\u003cspan class=\"CitationRef\"\u003e55\u003c/span\u003e] were subjected to a computer-assisted telephone interview, asking about post-COVID symptoms and the intensity: Information on 61 symptoms was collected based on the WHO Global SARS-COV-2 Clinical Platform Case Report Form for Post COVID condition. [\u003cspan class=\"CitationRef\"\u003e56\u003c/span\u003e] Participants were requested to report the symptoms (if any) and the duration in three time frames, 0\u0026ndash;3 months, 3\u0026ndash;6 months, or more than 6 months after infection. They were then classified as having post-COVID if they reported either (1) any new symptom onset after SARS-COV-2 infection, or (2) the persistence of pre-existing symptoms with increased severity after infection.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\n \u003ch2\u003eGutenberg Health Study\u003c/h2\u003e\n \u003cp\u003eThe Gutenberg COVID-19 study was based on the population sample of the Gutenberg Health Study (GHS), which provided additional information on the health of the participants before the pandemic. Established in 2007, GHS as a population-based prospective cohort study investigates diseases with high social impact, including cardiovascular disease, cancer, ophthalmological diseases, and immune and mental health conditions. With balanced sex representation (1:1 balanced), GHS enrolled over 20,000 participants aged 25\u0026ndash;85. On a quinquennial basis at a dedicated study center, extensive data acquisition encompasses parameters for intermediate clinical phenotypes and subclinical disease markers based on medical-technical examinations; medical, environmental, psychosocial, and lifestyle factors; biomaterial including but not limited to serum, Ethylenediaminetetraacetic acid (EDTA) plasma, and peripheral blood mononuclear cells (PBMC), urine, stool, and tear liquid; and genetic information. At the midpoint of the five-year interval, a computer-assisted telephone interview is carried out to survey the investigated diseases. The detailed methodologies employed for the data compilation were documented elsewhere [\u003cspan class=\"CitationRef\"\u003e57\u003c/span\u003e].\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\n \u003ch2\u003eGutenberg Post COVID Study\u003c/h2\u003e\n \u003cp\u003eSet up in 2021, the GPC was designed to investigate the long-term effects of SARS-COV-2 infection. The cohort incorporated 607 participants at the time of analysis, aged 20\u0026ndash;92 years, ensuring balanced sex representation (1:1). Participants underwent a comprehensive multidisciplinary clinical examination program over several days after the final diagnosis of post-COVID. This investigation included biobanking of various biospecimens for multi-omics analyses: blood fractions (DNA, RNA, serum, EDTA plasma, citrate plasma, and PBMCs), tooth pocket swabs, urine, tear liquid, stool samples, and routine laboratory. Throughout the examinations, a wide range of (sub)clinical phenotypes including but not limited to heart, lungs, brain, and abdomen was evaluated through diagnostic tools such as Holter electrocardiography, spirometry, smell testing, neuroimaging, and cognitive/psychological tests.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\n \u003ch2\u003eHigh-throughput proteomic profiling\u003c/h2\u003e\n \u003cp\u003eThe venous blood EDTA plasma samples were collected from GCS participants, aliquoted, and stored at -80\u0026deg;C within a dedicated biobank. They were then subjected to multiplex proximity extension assays (PEA), a state-of-the-art technology provided by Olink Bioscience (Uppsala, Sweden) using eight 384-plex panels. This strategic selection enabled the relative quantification of a repertoire of 2,945 unique proteins per individual. A detailed description of the Olink Explore 3072 methodology is provided in the \u003cstrong\u003eSupplementary Methods\u003c/strong\u003e. The measurements were reported in Normalized Protein eXpression (NPX) units on a log2 scale, where higher values indicate increased protein concentrations. The NPX data were then transformed (via comparison of identity, logarithm, square root, and negative inverse) to reduce the absolute value of the skewness of each protein. Finally, these transformed data were then winsorized to a radius of 6 absolute median deviations around the median.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\n \u003ch2\u003eData operationalization and statistical analysis\u003c/h2\u003e\n \u003cp\u003ePolygenic risk scores, which represent the aggregated effect of risk alleles for polymorphisms at multiple genomic loci, serves as a quantitative measure for assessing genetic susceptibility to multifactorial traits and disease [\u003cspan class=\"CitationRef\"\u003e58\u003c/span\u003e]. In this work, polygenic risk score were used as a proxy for investigating the association between defined genetic predispositions and the risk of PCS. Polygenic risk score was derived using the established equation:\u003c/p\u003e\n \u003cp\u003e\u003cimg src=\"data:image/png;base64,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\" width=\"217\" height=\"50\"\u003e\u003c/p\u003e\n \u003cp\u003ewhere the PRS for an individual (i) is a summation of the Genome-Wide Association Study coefficient (GWAS) estimates for each variant (j), denoted as (\u0026beta;\u003csub\u003ej\u003c/sub\u003e), each multiplied by the count of risk alleles present for each respective variant (G\u003csub\u003eij\u003c/sub\u003e). Additionally, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\rho\\:\\)\u003c/span\u003e\u003c/span\u003e\u003csub\u003ej\u003c/sub\u003e is the GWAS p-value for a variant (j) and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\rho\\:\\)\u003c/span\u003e\u003c/span\u003e\u003csub\u003eT\u003c/sub\u003e is the predetermined threshold. A detailed description of PRS estimation is provided in the \u003cstrong\u003eSupplementary Methods\u003c/strong\u003e.\u003c/p\u003e\n \u003cp\u003eDemographic and clinical characteristics of the study sample are presented using absolute counts and relative frequencies. The least absolute shrinkage and selection operator (LASSO)-regularized logistic regression model, with leave-one-out cross-validation (LOOCV), was employed to identify a core yet optimal set of proteins without redundant information. Correlation between symptoms was obtained by applying Pearson correlation to the symptom-specific protein score \u003cem\u003e(P\u003c/em\u003e\u003csub\u003e\u003cem\u003es\u003c/em\u003e\u003c/sub\u003e\u003cem\u003e)\u003c/em\u003e. The \u003cem\u003eP\u003c/em\u003e\u003csub\u003e\u003cem\u003es\u003c/em\u003e\u003c/sub\u003e was established as follows:\u003c/p\u003e\n \u003cp\u003e\u003cimg src=\"data:image/png;base64,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\" height=\"51\" width=\"116\"\u003e\u003c/p\u003e\n \u003cp\u003ewhere n is the number of selected proteins by the model, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\varvec{x}\\)\u003c/span\u003e\u003c/span\u003e is the measured value of the protein, and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\beta\\:\\)\u003c/span\u003e\u003c/span\u003e is the protein-related coefficient.\u003c/p\u003e\n \u003cp\u003eThe relationship between symptom specific protein signature correlations and their co-prevalence was investigated using correlation analysis.\u003c/p\u003e\n \u003cp\u003eLinear regression was utilized to investigate the association between pre-existing conditions and disease risk factors with symptom-specific protein scores. Similarly, linear regression was applied to investigate the relevance of symptom-oriented protein scores and objective clinical tests. The association between polygenetic risk scores and symptom manifestations was analyzed using linear regression. All analyses were adjusted for age and sex and were of exploratory nature with p-values (P) considered as continuous measure of statistical evidence. All statistical analyses were performed in R (version 4.3.1) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.r-project.org/\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e\n \u003ch2\u003eBioinformatic analysis\u003c/h2\u003e\n \u003cp\u003eTo augment the understanding of the results procured, bioinformatics-based methods were employed. Pathway enrichment analysis was conducted using Metascape (v3.5.20240901) [\u003cspan class=\"CitationRef\"\u003e59\u003c/span\u003e] to elucidate the pathways associated with symptom clusters. The \u003cem\u003eigraph\u003c/em\u003e R package (version 2.1.4) [\u003cspan class=\"CitationRef\"\u003e60\u003c/span\u003e] was utilized to visually depict the link between the pathways and the symptom clusters.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"Declarations","content":"\u003ch2\u003eConflict of interest\u003c/h2\u003e \u003cp\u003eThere are no conflicts of interest reported regarding this work. Outside the submitted work, PSW reports grants from Bayer AG, non-financial grants from Philips Medical Systems, grants and consulting fees from Boehringer Ingelheim, grants and consulting fees from Novartis Pharma, grants and consulting fees from Sanofi-Aventis, grants, consulting and lecturing fees from Bayer Health Care, grants and consulting fees from Daiichi Sankyo Europe, lecturing fees from Pfizer Pharma, lecturing fees from Bristol Myers Squibb, consulting fees from Astra Zeneca, consulting fees and non-financial support from Diasorin and non-financial support from IEM. PSW is funded by the Federal Ministry of Education and Research (BMBF 01EO1503), Ministry of Science and Health of the State of Rhineland Palatinate (MWG RLP 724\u0026thinsp;\u0026minus;\u0026thinsp;0010#2021/0030-1501), and the Federal Institute for Occupational Safety and Health (BAuA, F2447 / 537727) outside the present work. PSW is principal investigator of the German Center for Cardiovascular Research (DZHK) and principal investigator of the DIASyM research core (BMBF DIASyM research core (BMBF 161L0217A, 031L0217A). AKS received financial support by Abbvie, Bayer, Heidelberg Engineering, Novartis, and Santen outside the topic of this work. The remaining authors report no conflict of interest.\u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eThe study was funded by the European Regional Development Fund and the Ministry of Science and Health of the State of Rhineland-Palatinate (EFRE/REACT-EU, Grant No. 84007232 and No. 84009735); by the Federal Ministry of Education and Research (\u0026ldquo;EPIC-AI\u0026rdquo;, Grant No. 01EQ2401A) for the topic of evaluation of new approaches to data analysis and data sharing in long/post COVID-19 research; by the ReALity Initiative of the Life Sciences of the Johannes Gutenberg University Mainz for the establishment of a cell bank; and by the National University Medicine Research Network on Covid-19 (\u0026rdquo;NaFoUniMedCovid19\u0026rdquo;, Grant No. 01KX2021) B-FAST for the topics \u0026ldquo;poor living conditions\u0026rdquo; and \u0026ldquo;working conditions\u0026rdquo; and their association with COVID-19 in the population. Biomaterials were stored at the BioBank Mainz of the University Medical Center Mainz.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e \u003cp\u003eWe are indebted to all study participants of the Gutenberg COVID-19 study for their contribution. We thank all study staff and researchers involved in the planning and conduct of the study.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eGroff D et al (2021) Short-term and long-term rates of postacute sequelae of SARS-CoV-2 infection: a systematic review. JAMA Netw open 4:e2128568\u0026ndash;e2128568\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFernandez-de-Las-Pe\u0026ntilde;as C et al (2023) Persistence of post-COVID symptoms in the general population two years after SARS-CoV-2 infection: A systematic review and meta-analysis. J Infect\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOronsky B et al (2023) A review of persistent post-COVID syndrome (PPCS). Clin Rev Allergy Immunol 64(1):66\u0026ndash;74\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSoriano JB et al (2022) A clinical case definition of post-SARS-COV-2 condition by a Delphi consensus. Lancet Infect Dis 22(4):e102\u0026ndash;e107\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTsampasian V et al (2023) Risk factors associated with post\u0026thinsp;\u0026ndash;\u0026thinsp;SARS-COV-2 condition: a systematic review and meta-analysis. \u003cem\u003eJAMA internal medicine\u003c/em\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDavis HE et al (2023) Author correction: long COVID: major findings, mechanisms and recommendations. Nat Rev Microbiol 21(6):408\u0026ndash;408\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBatiha G, El-Saber et al (2022) Pathophysiology of Post-COVID syndromes: a new perspective. Virol J 19(1):158\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEckhardt M et al (2020) A systems approach to infectious disease. Nat Rev Genet 21(6):339\u0026ndash;354\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDemichev V et al (2021) A time-resolved proteomic and prognostic map of SARS-COV-2. Cell Syst 12(8):780\u0026ndash;794\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCaptur G et al (2022) Plasma proteomic signature predicts who will get persistent symptoms following SARS-CoV-2 infection. \u003cem\u003eEBioMedicine 85\u003c/em\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu T, Wang Y, Guo T (2022) Multi-omics in SARS-COV-2: Seeing the unseen but overlooked in the clinic. Cell Rep Med \u003cem\u003e3\u003c/em\u003e.3\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu M et al (2020) Transcriptional and proteomic insights into the host response in fatal SARS-COV-2 cases. \u003cem\u003eProceedings of the National Academy of Sciences 117\u003c/em\u003e.45 : 28336\u0026ndash;28343\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVillar M et al (2021) Characterization by quantitative serum proteomics of immune-related prognostic biomarkers for SARS-COV-2 symptomatology. Front Immunol 12:730710\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e1Su, Y et al (2022) Multiple early factors anticipate post-acute SARS-COV-2 sequelae. \u003cem\u003eCell 185\u003c/em\u003e.5 : 881\u0026ndash;895\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKroenke, Kurt RL, Spitzer, Janet BW (2001) The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med 16(9):606\u0026ndash;613\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eenkins C, David et al (1988) A scale for the estimation of sleep problems in clinical research. J Clin Epidemiol 41(4):313\u0026ndash;321\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNasreddine ZS et al (2005) The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment. J Am Geriatr Soc 53(4):695\u0026ndash;699\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang J-M (2007) Cytokines, inflammation, and pain. Int Anesthesiol Clin 45(2):27\u0026ndash;37\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCavalli G et al (2021) Interleukin 1α: a comprehensive review on the role of IL-1α in the pathogenesis and treatment of autoimmune and inflammatory diseases. Autoimmun rev 20(3):102763\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGregoriou S et al (2010) Cytokines and other mediators in alopecia areata. \u003cem\u003eMediators of inflammation 2010\u003c/em\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMorgun EI, Ekaterina A (2020) Vorotelyak. Epidermal stem cells in hair follicle cycling and skin regeneration: a view from the perspective of inflammation. Front cell Dev biology 8:581697\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVan Lenten, Brian J et al (2001) High-density lipoprotein loses its anti-inflammatory properties during acute influenza a infection. Circulation 103 18:2283\u0026ndash;2288\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eParween F, Rinkoo Devi Gupta (2022) Insights into the role of paraoxonase 2 in human pathophysiology. J Biosci 47(1):4\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMaes M et al (2011) Increased plasma peroxides as a marker of oxidative stress in myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS). Med Sci monitor: Int Med J experimental Clin Res 17(4):SC11\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSmall DM, John Prescott (2005) Odor/taste integration and the perception of flavor. Exp Brain Res 166:345\u0026ndash;357\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDumas L-E et al (2024) Impact of post-SARS-COV-2 olfactory disorders on quality of life, hedonic experiences and psychiatric dimensions in general population. BMC Psychiatry 24(1):111\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePark H et al (2021) The association between olfactory function and cognitive impairment in older persons with cognitive impairments: a cross-sectional study. \u003cem\u003eHealthcare. Vol. 9\u003c/em\u003e. No. 4. MDPI\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLebedeva NN et al (2020) Correlation between Measures of Anxiety and Heart Rate Variability in Men and Women with Affective Disorders. Neurosci Behav Physiol 50:743\u0026ndash;749\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSteinberger DC, Deanna M (2023) Barch. Investigating the link between depression, cognition, and motivation in late childhood. Child Psychiatry Hum Dev 54(2):570\u0026ndash;581\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLv Y et al (2022) Bioinformatics and systems biology approach to identify the pathogenetic link of Long COVID and Myalgic Encephalomyelitis/Chronic Fatigue Syndrome. Front Immunol 13:952987\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJha P, Kumar et al (2021) Gene expression profiling reveals the shared and distinct transcriptional signatures in human lung epithelial cells infected with SARS-CoV-2, MERS-CoV, or SARS-CoV: potential implications in cardiovascular complications of SARS-COV-2. Front Cardiovasc Med 7:623012\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMorris G et al (2018) Leaky brain in neurological and psychiatric disorders: Drivers and consequences. Australian New Z J Psychiatry 52(10):924\u0026ndash;948\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eL\u0026oacute;pez-Hern\u0026aacute;ndez Yamil\u0026eacute; et al (2023) The plasma metabolome of long COVID patients two years after infection. Sci Rep 13(1):12420\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBader SM et al (2022) Programmed cell death: the pathways to severe SARS-COV-2? Biochem J 479(5):609\u0026ndash;628\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThiemann S, Linda G (2011) Baum. The road less traveled: regulation of leukocyte migration across vascular and lymphatic endothelium by galectins. J Clin Immunol 31:2\u0026ndash;9\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLouveau A, Mesquita SD, Kipnis J (2016) Lymphatics in neurological disorders: a neuro-lympho-vascular component of multiple sclerosis and Alzheimer\u0026rsquo;s disease? \u003cem\u003eNeuron 91\u003c/em\u003e.5 : 957\u0026ndash;973\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLibby P, L\u0026uuml;scher T (2020) COVID-19 is, in the end, an endothelial disease. Eur heart J vol 41(32):3038\u0026ndash;3044\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDenorme F, Vanhoorelbeke K, Simon F, De Meyer (2019) von Willebrand factor and platelet glycoprotein Ib: a thromboinflammatory axis in stroke. Front Immunol 10:495067\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRubenfire M et al (1986) Platelet hyperaggregability in patients with chest pain and angiographically normal coronary arteries. Am J Cardiol 57:657\u0026ndash;660\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAllendes FJ et al (2023) Cardiovascular and autonomic dysfunction in long-COVID syndrome and the potential role of non-invasive therapeutic strategies on cardiovascular outcomes. Front Med 9:1095249\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKato A, Schleimer RP (2022) Bleier. Mechanisms and pathogenesis of chronic rhinosinusitis. J Allergy Clin Immunol 149:1491\u0026ndash;1503\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eParente L, Solito E (2004) Annexin 1: more than an anti-phospholipase protein. Inflamm Res 53:125\u0026ndash;132\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYounger J, Michael et al (2004) A foldable CFTR∆F508 biogenic intermediate accumulates upon inhibition of the Hsc70\u0026ndash;CHIP E3 ubiquitin ligase. J Cell Biol 167(6):1075\u0026ndash;1085\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang Y, Mandelkow E (2016) Tau in physiology and pathology. Nat Rev Neurosci 17(1):22\u0026ndash;35\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGriffin W, Sue T et al (2006) Interleukin-1 mediates Alzheimer and Lewy body pathologies. J Neuroinflamm 3:1\u0026ndash;9\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKarl\u0026eacute;n A et al (2009) Nogo receptor 1 regulates formation of lasting memories. Proceedings of the National Academy of Sciences 106.48 : 20476\u0026ndash;20481\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTan BK, Jyn et al (2022) Prognosis and persistence of smell and taste dysfunction in patients with SARS-CoV-2: meta-analysis with parametric cure modelling of recovery curves. \u003cem\u003ebmj 378\u003c/em\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAjmani GS et al (2017) Smoking and olfactory dysfunction: a systematic literature review and meta-analysis. Laryngoscope 127(8):1753\u0026ndash;1761\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDing J et al (2023) Association between inflammatory biomarkers and venous thromboembolism: a systematic review and meta-analysis. Thromb J 21(1):82\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlhassan S et al (2017) Clinical presentation and risk factors of venous thromboembolic disease. Crit Care Nurs Q 40(3):201\u0026ndash;209\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOgungbe O et al (2022) Cardiac postacute sequelae symptoms of SARS-CoV-2 in community-dwelling adults: cross-sectional study. Open Heart 9(2):e002084\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRahimi K (2020) Self-palpation for detection of paroxysmal atrial fibrillation: Much noise with little signal. PLoS Med 17(3):e1003098\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIadecola C et al (2016) Impact of hypertension on cognitive function: a scientific statement from the American Heart Association. \u003cem\u003eHypertension 68\u003c/em\u003e.6 : e67-e94\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAung N et al (2020) Causal inference for genetic obesity, cardiometabolic profile and SARS-COV-2 susceptibility: a Mendelian randomization study. Front Genet 11:586308\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBaumk\u0026ouml;tter R et al (2025) Risk tools for predicting long-term sequelae based on symptom profiles after known and undetected SARS-CoV-2 infections in the population. \u003cem\u003eEuropean Journal of Epidemiology\u003c/em\u003e, : \u003cem\u003ein revision\u003c/em\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWorld Health Organization (\u0026lrm;2021)\u0026lrm;. Global COVID-19 clinical platform case report form (\u0026lrm;CRF)\u0026lrm; for post COVID condition (\u0026lrm;post COVID-19 CRF)\u0026lrm;, 9 February 2021. World Health Organization\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWild P et al (2012) The Gutenberg Health Study. Fed Health Gaz 55:824\u0026ndash;830\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArnold N, Koenig W (2021) Polygenic risk score: clinically useful tool for prediction of cardiovascular disease and benefit from lipid-lowering therapy? Cardiovasc Drugs Ther 35(3):627\u0026ndash;635\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou Y et al (2019) Metascape provides a biologist-oriented resource for the analysis of systems-level datasets. Nat Commun 10(1):1523\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCs\u0026aacute;rdi G, Nepusz T et al (2024) igraph: Network Analysis and Visualization in R\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003eTable 1 is available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"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":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6024310/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6024310/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe emergence of post-COVID sequelae (PCS) represents a global challenge. However, understanding of biological mechanisms and the definition of quantifiable risk factors remains limited. This study harnessed the power of predictive machine learning models to explore the potential of proteomics in predicting individual post-COVID symptoms and their collective manifestation as PCS. The analysis utilized a panel of approximately 2900 proteins measured in 495 COVID-19 patients. The study identified 235 unique proteins associated with 21 distinct post-COVID symptoms. Symptoms more closely linked by similar protein patterns tended to co-occur more frequently in patients. Six symptom clusters with distinct molecular pathway signatures were uncovered, with metabolic and inflammatory pathways prominently involved across several clusters. The relevance of the specific protein signatures for post-COVID symptoms could be demonstrated and explored by objective, quantifiable clinical tests, including cognitive and somatic assessments, and underlined their relevance. Data from various modalities, including pre-existing conditions, disease risk factors and genetic susceptibility, revealed relevant relations that may contribute to PCS heterogeneity. This work underscores the complex and multifaceted nature of post-COVID symptoms. It emphasizes the need for systematic and more specific approaches to facilitate the development of targeted therapies and treatment strategies.\u003c/p\u003e","manuscriptTitle":"Proteomic Atlas of Post-COVID Sequelae","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-03-05 10:15:59","doi":"10.21203/rs.3.rs-6024310/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"265f0ee0-2174-4456-aae2-4db72db71cc0","owner":[],"postedDate":"March 5th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":44526539,"name":"Health sciences/Diseases/Infectious diseases/Viral infection"},{"id":44526540,"name":"Biological sciences/Computational biology and bioinformatics/Protein function predictions"}],"tags":[],"updatedAt":"2025-03-05T10:15:59+00:00","versionOfRecord":[],"versionCreatedAt":"2025-03-05 10:15:59","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6024310","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6024310","identity":"rs-6024310","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}