Overview

VCF files with GL and PL work. Run several burnins to optimize the widths of the proposal functions.

Input

• VCF file, created by e.g. ATLAS task major/minor
• txt file (optional): e.g. samplesPopulations.txt

This file is a user-created .txt file containing the samples to be used and their population affiliation. Different allele counts will be estimated for different populations

Example:

Output

• A text file with suffix _inbreedingMCMC.txt: This file contains the samples from the posterior distribution of the parameters F, gamma, pi, and a given number of allele frequencies.
• A text file with suffix _inbreedingMCMC_posteriors.txt.gz: This file contains information about the posteriors of the parameter values. One of these files is written to disk every 1000 iterations.

Usage Example

./atlas task=inbreeding vcf=example.vcf.gz

Specific Arguments

Reading vcf:

• samples: specify samples to be used and their population affiliation
• limitLines: amount of lines to be read from VCF file
• minDepth: only store sites with minimum depth
• minSamplesWithData: only store sites with minimum number of samples. Default = 1
• minMAF: only store sites where initial estimate of allele frequency is larger or equal to minMAF. Default = 0.0
• minVariantQuality: only store sites with minimum variant quality
• reportFreq: after how many lines the reading progress is printed to the terminal. Default = 10000.
• epsF: epsilon for EM algorithm to estimate allele frequencies. Default = 0.0001

MCMC:

• writeBurninToFile: write values samples during burnins to the results file.
• trueAlleleFreq: provide file with true allele frequencies (for debugging)
• thinning: only print every \(n^{th}\) iteration to the results file. Default = 1 (all)
• numBurnins: number of burn-in rounds should that should be run. Default = 1
• burninLength: number of iterations in each burn-in round. Default = 1000
• numIter: number of iterations in the MCMC run
• probMovingToModelNoF: probability of proposing move to model \(M_0\)