Inbreeding Coefficients

Method for calculating inbreeding coefficients using ngsF. See the paper for full details on the method.

To run this method you should run:

bash ./scripts/F.sh ./scripts/F_TAXON.conf

with the proper taxon name filled in.

Input files

Scripts

Script filename: F.sh example configuration file: F_TAXON.conf

Necessary input files

• data/TAXON_samples.txt bam list

Output files

• results/TAXON.approx_indF results from approximate EM calculation--used as a prior for full EM
• results/TAXON.approx_indF.pars binary file
• results/TAXON.indF inbreeding coefficient results
• results/TAXON.indF.pars binary file

Final results are in the *.indF file. Files labeled *.approx.indF are rough estimate files to speed up analysis in the full EM step of the calculation.

Mandatory F_TAXON.conf variables

• TAXON name of taxon used
• TAXON_LIST filename of bam list
• ANC_SEQ filename of ancestral sequence used this should already be in your common variables file
• REF_SEQ filename of reference sequence used this should already be in your common variables file

Optional F_TAXON.conf variables

• MIN_BASEQUAL minimum base quality (default=20)
• GT_LIKELIHOOD estimate genotype likelihoods (default=1)
• DO_GLF type of GL output filetype (default=3)
• MIN_MAPQ minimum base mapping quality (default=30)
• N_CORES number of cores to use (default=32) make sure to adjust this according to what your computer or cluster allows. This may need to be adjusted depending on the memory requirements of your data
• DO_MAJORMINOR estimate major/minor alleles (default=1)
• DO_MAF calculate per site frequencies (default=1)
• SNP_PVAL=1e-6 p-value limit for calling SNPs (default=1e-6) removing invariant sites with a light SNP call greatly decreases computational time and numerical instability
• OVERRIDE if true, will recalculate files that already exist (default=false)
• SEED=12345 set for approx. EM--do not change
• MIN_EPSILON=1e-9 epsilon value for EM converging. A smaller number will greatly increase computational time. 1e-7 or smaller recommended for proper converging, but adjust if computational time too intensive (default=1e-9)