Regulation of protein abundance in genetically diverse mouse populations

Abstract

Proteins constitute much of the structure and functional machinery of cells, forming signaling networks, metabolic pathways, and large multi-component complexes. Protein abundance is regulated at multiple levels spanning transcription, translation, recycling, and degradation to maintain proper balance and optimal function. To better understand how protein abundances are maintained across varying genetic backgrounds, we analyzed liver proteomes of three genetically diverse mouse populations. We observe strong concordance of genetic and sex effects across populations. Differences between the populations arise from the contributions of additive, dominance, and epistatic components of heritable variation. We find that the influence of genetic variation on proteins that form complexes relates to their co-abundance. We identify effects on protein abundance from mutations that arose and became fixed during breeding and can lead to unique regulatory responses and disease states. Genetically diverse mouse populations provide powerful tools for understanding proteome regulation and its relationship to whole-organism phenotypes.

Multi-Omic Profiling of the Liver Across Diets and Age in a Diverse Mouse Population

“…cis-QTLs—are more likely to have more significant transcript–protein relationships ( Figure 2F ). QTLs indicate a causal impact of DNA sequence variants inside (or adjacent to) a gene which cause varying transcript expression levels (cis-eQTL) or protein levels (cis-pQTL), and these tend to be highly robust and reproducible [30]. The knowledge of which genes have cis-QTLs can also provide predictive information across expression type: transcripts that have strong cis-eQTLs (logarithm of the odds (LOD) ≥ 4) correlate substantially better with their protein (rho ∼ 0.27).…”

A Bayesian model selection approach to mediation analysis

“…In order to illustrate how bmediatR works in QTL mapping applications, we analyzed previously reported liver proteomics data from 192 DO mice [5,13]. We first illustrate how multi-state haplotypes help to identify the causal driver of a distal pQTL for Snx4.…”

“…The examples related to the distal pQTL of Snx4 and Tubg1 were previously identified [13]. QTL were mapped using the following linear mixed effect model:…”

“…Samples of liver tissue were collected and processed for quantitative massspectrometry as previously described [5]. Estimation and normalization of the protein abundance data from component quantitative peptide data and subsequent QTL analysis have been previously described [13]. Genetic mapping was based on final quantities output from the rank-based inverse normal transformation (RINT) [53].…”

“…Applications of mediation analysis to large scale genetic and molecular profiling data have used modified versions of the traditional tests described above. Approximations to CS have been used in the multi-parent Collaborative Cross (CC [12,6,13]) and Diversity Outbred mouse populations (DO [14,5,15,16]). These studies identified QTL for gene expression (eQTL), protein abundance (pQTL), and chromatin accessibility (cQTL).…”