This GitHub repository contains the scripts that were used to generate the analyses and figures in:
Beichman, Robinson, Lin, Nigenda-Morales, Moreno Estrada, & Harris. "Evolution of the mutation spectrum across a mammalian phylogeny," Molecular Biology & Evolution (in revision)
Figure 1 from Beichman et al., showing an overview of the analytical pipeline
The main code directories are as follows, each of which has its own detailed readme file:
- 0_dataProcessingandPolarizationScripts/: scripts used to assign ancestral allele states
- 1_GenerateMutationSpectraUsingMutyper/: a pipeline used to generate mutation spectra for every species in the dataset
- 2_PCA_onSpectra/: scripts used to generate principal component analyses (PCA) based on mutation spectra
- 3_PhyloSignalOfSpectra_and_Enrichments/: scripts used to carry out analysis of phylogenetic signal of the mutation spectrum, correlation of the spectra with possible confouners, and enrichments/depletions of particular k-mers
- 4_PhyloSignalOfConfounders/: scripts used to measure the phylogenetic signal of technical and biological confounders
- 5_MutationSignatureFitting_inSigfit/: scripts used to carry out mutational signature fitting analyses
- 6_ComparingToMouseDNMs/: scripts used to carry out additional comparisons of mouse-wolf 1-mer spectra similarities using two alternative datasets
Each directory has its own detailed readme file containing script-specific information (input and output files, overviews of functions, meanings of parameters, etc.), and scripts are heavily commented. There are extensive details on the analyses in the SI Methods section of Beichman et al.
- Figure 2 (Principal component analyses): 2_PCA_onSpectra/
- Figures 3, 4, & 5: 3_PhyloSignalOfSpectra_and_Enrichments/
- Figures 6 & 7: 5_MutationSignatureFitting_inSigfit/
Data files from the paper are on Dryad (DRYAD LINK) with an extensive readme describing them. Smaller auxiliary files have been placed in each script's GitHub directory for convenience.
You would need to update paths to input files on your own system inside script if re-running scripts. We recommend that R scripts be run in RStudio step-by-step to see what each code chunk is doing. All code is presented as-is (exactly as it was used to generate the analyses of the paper), so there may be extraneous comments and analyses present within them.