This repository contains the R and stan code used to perform the analysis discussed in the manuscript "Metabotyping as a Stopover in Genome-to-Phenome Mapping" by Pubudu Handakumbura et al. published in Scientific Reports. The data needed to run this code is available from osti.gov here.
The contents of the Code folder are as follows:
- bayes_allometric.R - This code fits the allometric model via Bayesian hierarchical modelling (BHM) using
stan - BHM_allom_lscale.stan - This script defines the Bayesian hierarchical model
- Create_figures_2_and_Ex2.R - Code necessary to create Figure 2 and Extended Data Figure 2.
- Create_RF_Imp_Plot.R - This script collects the random forest results (created in 'rf_biomass_pred.R') and plots the features based on their feature importance
- Dendrogram_Comparison.R - This script compares the above and belowground dendrograms using Baker's gamma then performs a permutation test to assess the statistical significance of the observed Baker's gamma value
- metab_analysis.R - This script uses the
MSomicsSTATpackage to analyze the raw metabolite abundance data. This script creates the heat maps in Figure 4, Extended Data Figure 4 and the scrollable heatmaps in this shiny app - PCA_PERMONOVA.R - This script performs a principal component analysis and PERMONOMVA
- pheno_funs.R - This script contains a variety of functions used in the other scripts
- rf_biomass_pred.R - This script fits the random forest model to predict biomass as a function of metabolite abundances