Assigning Co-Regulated Human Genes and Regulatory Gene Clusters

Abstract

Elucidating the role of genetic variation in the regulation of gene expression is key to understanding the pathobiology of complex diseases which, in consequence, is crucial in devising targeted treatment options. Expression quantitative trait locus (eQTL) analysis correlates a genetic variant with the strength of gene expression, thus defining thousands of regulated genes in a multitude of human cell types and tissues. Some eQTL may not act independently of each other but instead may be regulated in a coordinated fashion by seemingly independent genetic variants. To address this issue, we combined the approaches of eQTL analysis and colocalization studies. Gene expression was determined in datasets comprising 49 tissues from the Genotype-Tissue Expression (GTEx) project. From about 33,000 regulated genes, over 14,000 were found to be co-regulated in pairs and were assembled across all tissues to almost 15,000 unique clusters containing up to nine regulated genes affected by the same eQTL signal. The distance of co-regulated eGenes was, on average, 112 kilobase pairs. Of 713 genes known to express clinical symptoms upon haploinsufficiency, 231 (32.4%) are part of at least one of the identified clusters. This calls for caution should treatment approaches aim at an upregulation of a haploinsufficient gene. In conclusion, we present an unbiased approach to identifying co-regulated genes in and across multiple tissues. Knowledge of such common effects is crucial to appreciate implications on biological pathways involved, specifically when a treatment option targets a co-regulated disease gene.

Transcription factor expression is the main determinant of variability in gene co‐activity

“…Overall, co-activity scores within defined co-activity domains appeared stable across individuals and were reflective of local chromatin states, conforming to observations that topological and compartment domains have a tendency to remain consistent across cell types and individuals in a population (Rao et al, 2014;Gorkin et al, 2019). Furthermore, the co-regulatory potential within a domain was manifested by shared eQTLs having an impact on multiple genes within the same domain, supported by observations of eGene pairs situated in close proximity within the genome (Strunz et al, 2021). We found regions within co-activity domains that displayed significant variability in co-activities to be particularly interesting, as their analysis allowed for a deeper understanding of the underlying mechanisms driving their variability.…”

Transcription factor expression is the main determinant of variability in gene co-activity

“…Overall, co-activity scores within defined co-activity domains appeared stable across individuals and were reflective of local chromatin states, conforming to observations that topological and compartment domains have a tendency to remain consistent across cell types and individuals in a population (Gorkin et al , 2019; Rao et al , 2014). Furthermore, the co-regulatory potential within a domain was manifested by shared eQTLs having an impact on multiple genes within the same domain, supported by observations of eGene pairs situated in close proximity within the genome (Strunz et al , 2021). We found regions within co-activity domains that displayed significant variability in co-activities to be particularly interesting, as their analysis allowed for a deeper understanding of the underlying mechanisms driving their variability.…”

Vitronectin and Its Interaction with PAI-1 Suggests a Functional Link to Vascular Changes in AMD Pathobiology

“…GTEx data processing protocols and filtering for European individuals are detailed in Ref. [ 21 ]. Overall, 556 whole blood samples were analyzed.…”

“…Linear regression models were applied to analyze the relationship of VTN and PAI-1 mRNA gene expression in ( A ) whole blood and ( B ) retinal tissue. The 556 whole blood samples were derived from European donors in the GTEx dataset and underwent several normalization steps, as described in [ 20 , 21 ]. The analysis in retina was based on 311 healthy retina samples, as described in [ 22 ].…”