Identify the heterogeneity and commonality of chronic overlapping pain conditions (COPCs) through phenotypic and genomic perspectives

点击翻译按钮获取中文摘要

PROJECT SUMMARY/ABSTRACT With an estimated prevalence of 20% in American adults and annual costs above $500 million, chronic pains pose a high toll to public health. Many patients developed chronic overlapping pain conditions (COPCs), where craniofacial pains like temporomandibular disorder (TMD) represent a unique component that co-occurs frequently with other chronic pains including irritable bowel syndrome (IBS). Not only do the comorbidities complicate pain management, but the etiology of COPCs remains unclear. Heterogeneity exists within and across COPCs, so mechanism-based classification schemes are needed to identify safe and effective therapies for distinct subgroups of patients. However, existing research surrounding COPC taxonomy has not fully integrated phenotypic and genotypic data at the population level. Reliance on disease-specific study cohorts seriously limited the sample size and diversity. On the other hand, informatics and data science have advanced secondary use of biomedical data, presenting a strong alternative to hypothesis-driven, controlled studies. We propose to elicit COPC subgroups by mining three distinct population-based clinical datasets and imputing the biological underpinnings of co-occurring COPCs by using functional genomic knowledge bases. Our approach consists of two aims: 1) Identify COPC subgroups and other commonly associated phenotypes from rich longitudinal clinical data and notes, including over one million patients in the Rochester Epidemiology Project. The clinical datasets will be computationally screened for COPCs and other co-occurring phenotypes based on diagnosis codes and natural language processing. We will identify statistically significant COPC comorbidities and progression trajectories using novel and tailored statistics. The discovered trajectories will be clustered into subgroups using cutting-edge graph clustering algorithms. The patients will be assigned to the best matched subgroups, for which additional phenotypic characteristics of each group will be determined by least absolute shrinkage and selection operator. 2) Impute biological underpinnings for comorbid COPCs by integrating phenotypic, genetic, and genomic data in biobanks and biorepositories. We will conduct genome-wide and phenome-wide association studies based on the diagnoses and genotypes from the UK Biobank and All of Us Research Program, leading to identification of additional genotypes that are associated with the COPCs and beyond those in the NHGRI-EBI GWAS catalog. We will apply our information-theoretic framework to impute the functional similarity and shared biological mechanisms across COPCs by using GTEx expression quantitative trait loci data and gene ontology annotations. The findings of shared mechanisms among COPCs will provide novel insight into the genetic factors, particularly in noncoding regions, and functional linkages that are pivotal to developing applications such as drug repurposing for COPCs. The two aims corroborate each other across the genome-phenome boundary to unveil interpretable subgroups that will advance precision medicine for COPCs.