Population-scale proteome variation in human induced pluripotent stem cells

Abstract

Human disease phenotypes are ultimately driven primarily by alterations in protein expression and/or function. To date, relatively little is known about the variability of the human proteome in populations and how this relates to variability in mRNA expression and to disease loci. Here, we present the first comprehensive proteomic analysis of human induced pluripotent stem cells (iPSC), a key cell type for disease modelling, analysing 202 iPSC lines derived from 151 donors, with integrated transcriptome and genomic sequence data from the same lines. We characterised the major genetic and non-genetic determinants of proteome variation across iPSC lines and assessed key regulatory mechanisms affecting variation in protein abundance. We identified 654 protein quantitative trait loci (pQTLs) in iPSCs, including disease-linked variants in protein coding sequences and variants with trans regulatory effects. These include pQTL linked to GWAS variants that cannot be detected at the mRNA level, highlighting the utility of dissecting pQTL at peptide level resolution.

Simultaneous quantification of mRNA and protein in single cells reveals post-transcriptional effects of genetic variation

“…Across several species, half or more of cis -eQTLs were reported to also affect protein abundance, suggesting high levels of agreement between mRNA and protein ( Albert et al, 2018 ; Albert et al, 2014b ; Hause et al, 2014 ; Skelly et al, 2013 ). However, other work found many cis -eQTLs whose effects on protein were smaller than those on mRNA, suggesting the existence of mechanisms that buffer protein levels against variation in mRNA abundance ( Battle et al, 2015 ; Ghazalpour et al, 2011 ; Mirauta et al, 2020 ). Trans -acting QTLs were reported to show even greater differences between mRNA and protein.…”

“…It is also unclear to what extent there are protein-specific pQTLs that influence protein abundance without affecting the mRNA of the same gene. Recent studies have begun to address these questions ( Battle et al, 2015 ; Cenik et al, 2015 ; Chick et al, 2016 ; Foss et al, 2011 ; Foss et al, 2007 ; Ghazalpour et al, 2011 ; Großbach et al, 2019 ; Mirauta et al, 2020 ; Picotti et al, 2013 ; Sun et al, 2018 ; Wu et al, 2013 ; Yao et al, 2018 ) and reported a wide range of observations.…”

“…Trans -acting QTLs were reported to show even greater differences between mRNA and protein. In mouse crosses and some yeast crosses, ( Mirauta et al, 2020 ) trans -eQTLs and trans -pQTLs mapped for the same genes showed very little overlap, suggesting that while many trans -eQTLs are buffered at the protein level, there are also many trans -pQTLs that affect proteins via mechanisms that do not affect mRNA ( Chick et al, 2016 ; Foss et al, 2011 ; Foss et al, 2007 ; Ghazalpour et al, 2011 ; Williams et al, 2016 ). While more recent work in yeast showed somewhat higher agreement between trans -eQTLs and trans -pQTLs ( Albert et al, 2018 ; Großbach et al, 2019 ), yeast crosses also revealed instances of ‘discordant’ trans -acting loci that affect both mRNA and protein of the same gene, but in opposite directions ( Albert et al, 2018 ; Albert et al, 2014b ; Foss et al, 2011 ; Großbach et al, 2019 ).…”

“…This challenge is not easy to overcome, because combined application of RNA-sequencing (RNA-seq) and mass spectrometry in thousands of matched samples remains prohibitively expensive. While recent studies in yeast ( Großbach et al, 2019 ) mice ( Chick et al, 2016 ; Williams et al, 2016 ), and human iPSC lines ( Mirauta et al, 2020 ) used RNA-seq and mass spectrometry on the same cultures or tissue samples, their sample sizes of less than 200 individuals limited QTL detection. As a result, the role of genetic variation on post-transcriptional processes remains unclear, especially for trans -acting variation.…”

Genetic analysis of blood molecular phenotypes reveals regulatory networks affecting complex traits: a DIRECT study

“…In conclusion, we estimated a large proportion of shared genetic regulation across gene expression and proteins. However, as reported by others before 5,16–18,22,23 , the genetic regulation of circulating proteins seems to involve additional protein specific regulatory processes that complicates the identification of genomic regions acting in both phenotypes.…”