Pruthi, Siddharth (2023) Proteomic profiling of skeletal muscle mitochondrial adaptations to exercise training through meta-analysis methods. Research Master thesis, Victoria University.

Abstract

Many studies have analysed the skeletal muscle mitochondrial proteome in relation to exercise training. These studies have identified adaptations within metabolic pathways and other pathways governing mitochondria form and function. With the adoption of high-throughput proteomics within the field, the depth of mitochondrial proteome coverage has increased many-fold within the last decade. Information on protein abundance changes following exercise training now exists for hundreds of mitochondrial proteins across multiple studies. Despite this greater data availability, no research has yet sought to compare, contrast, and aggregate the findings from these studies. Therefore, this research aims at a comprehensive analysis of adaptations of the skeletal muscle mitochondria proteome to exercise training using all available high-throughput proteomics data. Eight independent datasets were shortlisted to be included in the meta-analysis study. For each dataset, raw protein intensities were extracted for each sample from the respective proteinGroups file. Uniform filtration, imputation, and normalisation, if required, were applied to raw protein intensities and differential expression analysis was performed using a linear-modelling approach in DEqMS, a package built on top of limma and which accounts for peptide-identification information when adjusting the protein-wise variance estimator. A Random-Effects meta-analysis was performed on these aggregated differential expression results using the package metafor. Independently of these analyses, the effects of mitochondria-normalisation (in silico scaling of raw protein intensities to account for overall change in mitochondria protein abundance) on differential expression results was also explored. A total of 778 proteins were quantified across the eight datasets and the meta-analysis revealed 200 (K1 ≥ 5) differentially expressed proteins at FDR < 0.05. The capacity of the increased power of the meta-analysis design was also highlighted as approximately 20% of the significant proteins at the meta-analysis level were not significant in any individual dataset. The findings of the meta-analysis revealed an enrichment for terms related to the Electron Transport Chain, multiple metabolic pathways, and mitochondria biogenesis. While not demonstrating significant enrichment, multiple central proteins within pathways of ROS detoxification (SOD2, GSTK1, PRDX3), protein import and assembly (TIMM44, HSPA9, GRPEL1), and small molecule transport (SLC25A3, MTCH2) were identified as significantly upregulated. Exploratory analysis of heterogeneity scores (I2 %) of meta-analysis results suggest stronger effects of study variables (exercise-training protocol) on adaptations to oxidative metabolism proteins compared to the rest. Analysis of mitochondria-normalised data revealed different enrichment profiles depending upon exercise exposure of the associated study, suggesting the potential of this technique to help investigate the time-course of mitochondria proteome adaptations to exercise training independent of overall change in mitochondria content. This research aims to be the basis of a resource for the wider scientific community that can be used to generate novel hypotheses, select targets for further experimental validation, and investigate exercise-training induced adaptations at the level of the mitochondria protein network.

Search Google Scholar

Repository staff login