Projecting genetic associations through gene expression patterns highlights disease etiology and drug mechanisms

Abstract

Understanding how dysregulated transcriptional processes result in tissue-specific pathology requires a mechanistic interpretation of expression regulation across different cell types. It has been shown that this insight is key for the development of new therapies. These mechanisms can be identified with transcriptome-wide association studies (TWAS), which have represented an important step forward to test the mediating role of gene expression in GWAS associations. However, due to pervasive eQTL sharing across tissues, TWAS has not been successful in identifying causal tissues, and other methods generally do not take advantage of the large amounts of RNA-seq data publicly available. Here we introduce a polygenic approach that leverages gene modules (genes with similar co-expression patterns) to project both gene-trait associations and pharmacological perturbation data into a common latent representation for a joint analysis. We observed that diseases were significantly associated with gene modules expressed in relevant cell types, such as hypothyroidism with T cells and thyroid, hypertension and lipids with adipose tissue, and coronary artery disease with cardiomyocytes. Our approach was more accurate in predicting known drug-disease pairs and revealed stable trait clusters, including a complex branch involving lipids with cardiovascular, autoimmune, and neuropsychiatric disorders. Furthermore, using a CRISPR-screen, we show that genes involved in lipid regulation exhibit more consistent trait associations through gene modules than individual genes. Our results suggest that a gene module perspective can contextualize genetic associations and prioritize alternative treatment targets when GWAS hits are not druggable.

An efficient not-only-linear correlation coefficient based on machine learning

“…Recent theories, like the omnigenic model for complex traits [16,17], argue that these observations are explained by highly-interconnected gene regulatory networks, with some core genes having a more direct effect on the phenotype than others. Using this omnigenic perspective, we and others [19,20,23] have shown that integrating gene co-expression networks in genetic studies could potentially identify core genes that are missed by linear-only models alone like GWAS. Our results suggest that building these networks with more advanced and efficient correlation coefficients could better estimate gene co-expression profiles and thus more accurately identify these core genes.…”

“…The analysis of large RNA-seq datasets [10,11] can also reveal complex transcriptional mechanisms underlying human diseases [2,12,13,14,15]. Since the introduction of the omnigenic model of complex traits [16,17], gene-gene relationships are playing an increasingly important role in genetic studies of human diseases [18,19,20,21], even in specific fields such as polygenic risk scores [22]. In this context, recent approaches combine disease-associated genes from genome-wide association studies (GWAS) with gene co-expression networks to prioritize “core” genes directly affecting diseases [19,20,23].…”

“…Since the introduction of the omnigenic model of complex traits [16,17], gene-gene relationships are playing an increasingly important role in genetic studies of human diseases [18,19,20,21], even in specific fields such as polygenic risk scores [22]. In this context, recent approaches combine disease-associated genes from genome-wide association studies (GWAS) with gene co-expression networks to prioritize “core” genes directly affecting diseases [19,20,23]. These core genes are not captured by standard statistical methods but are believed to be part of highly-interconnected, disease-relevant regulatory networks.…”

A publishing infrastructure for AI-assisted academic authoring

“…Assessing the performance of an automated revision tool is not straightforward, since a review of a revision will necessarily be subjective. To mitigate this, we used three manuscripts of our own authorship (Table 1): the Clustermatch Correlation Coefficient (CCC) [9], PhenoPLIER [10], and Manubot-AI (this manuscript). CCC is a new correlation coefficient evaluated in transcriptomic data, while PhenoPLIER is a framework that comprises three different methods applied in the field of genetic studies.…”

BuDDI:Bulk Deconvolution with Domain Invarianceto predict cell-type-specific perturbations from bulk

“…There exist several generative methods to learn interpretable latent spaces that decompose the input single-cell expression profiles into relevant sources of variation. These methods can be directly trained to capture a specific source of variation [29][30][31][32][33][34][35] or post-hoc-interpreted after training [36][37][38][39][40] . Furthermore, there exist several methods to learn a latent space such that shifts within the latent space represent specific perturbation effects on an unobserved cell or cell type [4][5][6][7][8][9][10][11][12][13][14]28 .…”