Dataset_Govus_2025_JP
Information
Dataset to reproduce the statistical analysis undertaken in Govus et al., (2025) Acute Metabolic Phenotype Responses to Swimming Exercise of Different Intensities in Highly Trained Male and Female Swimmers, submitted for review to the Journal of Physiology.
Study Overview
This study analysed the changes in metabolomic profile in 16 highly trained male (n = 9) and female (n = 7) swimmers who performed each of the swimming trials below, separated by a 1-2 day break:
- Moderate domain trial: 5 × 400 m on 6’ at A1/A2 intensity
- Heavy domain trial: 3 × (8 × 100 m holding critical speed, 100 m recovery on 2’)
- Severe domain trial: 3 × (1 × 35 m dive max on 2’, 2 × 50 m dive max on 3’, 200 m recovery on 5’)
Project Team
- Dr Andrew Govus (La Trobe University, Principal Investigator)
- Dr Chloe Goldsmith (University of Western Sydney, Co-investigator)
- Dr Katie McGibbon (Queensland Academy of Sport, Co-investigator)
- Dr Lachlan Mitchell (Victorian Institute of Sport, Co-investigator)
- Dr Maria Kozlovskaia (University of Canberra, Co-Investigator)
- E/Prof David Pyne (University of Canberra, Co-Investigator)
- Dr Nathan Lawler (Australian National Phenome Centre, Murdoch Univeristy, Co-investigator)
Sample & Data Analysis
- Metabolomic data analysis: Blood plasma was analysed by NMR and LC-MS by Dr Nathan Lawler (Mudroch University) at the Australian National Phenome Centre.
- Bioinformatics: Bioinformatics for metabolomic data was performed by Andrew Govus (La Trobe University) and Dr Nathan Lawler (Murdoch University).
Project Funding
- Queensland Academy of Sport Innovation and Knowledge Excellence (SPIKE) ($25,000 AUD)
- University of Canberra Industry Collaborative Seed Grant ($25,000 AUD)
- La Trobe University School of Allied Health, Human Services & Sport Stategic Research Allocation ($10,000 AUD)
Bioinformatics Data Analysis
- Data analysis is peformed using the R statistical programming language.
Bioinformatics Approach - Metabolomics
- Data cleaning
- Inputation of missing data & outlier detection
- Supervised multivariate analysis: Orthogonal Partial Least Squeres Discriminant Analysis (OPLS-DA) on log2 fold change (post-exercise/pre-exercise) to compare the heavy and severe domain trials against the moderate domain trial
- Univariate analysis: Linear mixed models to compare each trial + eruption plot to visualise influential metabolites
Github link: https://github.com/NathanGLawler/Swimmer-s-Phenomics-Project
