Associated data and scripts for "Drainage-structuring of ancestral variation and a common functional pathway shape limited genomic convergence in natural high- and low-predation guppies"
Contact [email protected] with any issues
Raw data for this project is available at ENA: PRJEB43917 (Aripo, Madamas, Tacarigua) and PRJEB36704 (Guanapo and Oropouche). SNP calling was performed using the pipeline described here: https://github.com/josieparis/gatk-snp-calling
Many scripts reference five_aside_STAR, this is simply the shorthand name given to the main dataset (five HP and five LP populations, STAR = shorthand name for the genome assembly used).
VCF data is available from FigShare, doi:10.6084/m9.figshare.14315771, store in VCFs/. The main VCF here is used in the majority of analyses, and this version of the VCF has been phased by beagle/shapeit. For introgression analyses, we included additional available data for Poecilia wingei (see manuscript). Smoove-called SV VCFs are located in SV_and_coverage/data/.
General functions can be found in R/, these are sourced in various R scripts and the path may need updating to reflect their current path. These include important functions such as the liftover of scaffold 000094F_0 positions to the updated chr20.
general_data/ includes the fasta index for the male guppy genome, which is used by various scripts.
popmaps/ includes a popmap for all 10 populations + wingei individuals, popmaps for each of the five rivers, and a total popmap of all individuals in the main VCF.
Each subdirectory includes an environment within which to perform specific analyses. For example, BayPass/ includes scripts required to run BayPass aux model and calculate XtX within rivers starting with the VCF. Each analysis subdirectory includes a data/, outputs/, figs/, R/, scripts/, scripts/logs/, tables/ subdirectory with scripts based on relative paths assuming this structure. If any of these are not present, they should be made prior to running with mkdir. Some paths may need to be updated before scripts will run, although in most cases this should be case of resetting the MASTER path directory at the top of scripts as the majority of paths are relative.
The following directories include individual analyses, i.e. they are not dependent on each other to run:
dsuite/
fineSTRUCTURE/
BayPass/
RAxML/
SNAPP/
SV_and_coverage/
localPCA/
phasing/
Note: The scripts associated with fastsimcoal and SFS production were provided by Vitor Sousa and so are not presented here.
The three directories with *_analysis names should be run at the end, as these include scripts related to the three main sections of the manuscript and require as inputs the outputs from various analyses.