Computational Methods for Systems Genetic Analysis of Rare Polygenic Disorder

Idiopathic Pulmonary Fibrosis (IPF) inflicts a significant healthcare burden and high rate of mortality in the U.S. Current treatment options can slow the rate of lung function decline, but the five-year survival remains low, therefore there is clear unmet medical needs for novel therapeutics. Here, human genetics can catalyze drug

development process by identifying new biological targets and elucidating the underlying pathways. As a rare polygenic disorder, however, IPF poses challenges to existing disease gene-mapping strategies due to the extensive locus heterogeneity and difficulty of assembling massive sample sizes typical of common polygenic

disorders. This project aims to develop more effective gene-mapping methods for rare polygenic disorders such as IPF. Our approach is motivated by three complementary strategies for small genetic studies: (i) pleiotropy-informed SNP association tests, which can be extended to take advantage of the pleiotropy of

disease SNPs with gene expression traits and to further boost power by accounting for the network connectivity to known disease genes; (ii) Polygenic Risk Scores (PRS), which can be highly useful for rare disorders by capturing the effect of disease modifiers in the genetic background; and (iii) highly modular

network structure of disease genes, which can be leveraged to reduce genetic heterogeneity among cases. In Aim 1, by extending a pleiotropy-informed association test we had previously proposed, we will develop a new network model-based association test informed by pleiotropy to gene expression traits. We will apply this

method to publicly available IPF GWAS data and expression Quantitative Trait Loci (eQTL) of IPF-relevant tissues and cell populations. In Aim 2, we will develop a new rare-variant association test directly accounting for the contribution of genetic background using PRS. We will apply the new method to sequencing data of

~1,500 IPF cases and ~15,000 unaffected controls from CGS-PF and TOPMed studies. In Aim 3, we will identify novel IPF genes by leveraging the association between disease gene modules and comorbidities. Known IPF genes are clustered in multiple tightly inter-connected gene modules in biological networks, and

mutations disrupting each network modules cause a distinct set of comorbidities in IPF patients. We will leverage the modularity of IPF genes and comorbidity to find novel IPF genes in exome data of UK Biobank and MGB Biobank. In reverse, we also will test if genotypes of key IPF gene modules can inform the course of

comorbidity development in patients by inviting 10 CGS-PF study participants to a reverse genetics study. Ultimately, the findings from these studies will uncover novel genes and pathways underlying IPF and develop new computational strategies generally applicable to rare polygenic disorders.