5.3 call

Calling genotypes

call generates a VCF file that is used to encode genetic variant sites and genotypes. See VCFCompare for more information on VCF file format.call can make use of different callers to produce a VCF file. Some are maximum likelihood callers, like the “MLE” caller, and others are Bayesian callers, like “Bayesian” and “allelePresence”. For the Bayesian callers, we define the prior based on theta and the allele frequencies (see below for more details). The parameters of the priors can either be estimated in a first step and then in a second step assumed to be known when making the genotype call, or they can be fixed to a value.

5.3.1 Parameters

5.3.1.1 Input

--bam Input_bam_file.bam Input bam file
--fasta Input_refrence_genome_file.fasta reference genome FASTA file

5.3.1.2 Specific

Parameter Description Default
--method caller_type caller type for variant calling. Following options are available : MLE, Bayesian, allelePresence, randomBase, majorityBase MLE
--priors prior_type Prior for the callers allelePresence and Bayesian. theta and base frequencies estimated
--fixedTheta integer_value provide fixed theta value instead of estimating it for every window.
--equalBaseFreqs assume all base frequencies to be 0.25 instead of estimating them for every window.
defaultTheta integer_value provide a fixed theta value to be used for all windows for which theta can not be estimated due to lack of data (algorithm does not converge).
--infoFields DP add depth to vcf-file
--formatFields format_fields Print the VCF format fields that are specified. Available options are: GT: genotype string, DP: sequencing depth, GQ: genotype quality, AD: allelic depths for all alleles in call in order listed, AP: Phred-scaled allelic posterior probabilities for the four alleles A, C, G and T, GL: normalized genotype likelihoods, PL: phred-scaled normalized genotype likelihoods, GP: Genotype posterior probabilities (phred-scaled), AB: alleleic imbalance, AI: Binomial probability of allelic imbalance if Hz site. All arguments must be provided as a string, separated by comma (,). e.g. : --formatFields GT,DP,AD, GQ, PL. GT,DP,AD,GQ,PL
--printAll Print all sites, also invariant ones. Only sites with data.
--noAltIfHomoRef Specify to not print alternative alleles if call is homozygous reference, but still use them to calculate call quality. Print the most likely alternative allele even if the call is homozygous reference
--noTriallelic Only allow one alternative allele. Allow for genotypes with two alternative alleles.
--noCallsViolatingBest Do not call genotypes from known alleles that conflict with best call across all genotypes. Call genotypes from known alleles even if they differ from best call across all genotypes.
--sampleName name Define a sample name for the header of the vcf file. prefix specified with the out parameter
--alleles Limit calls to sites with known alleles. Will call without prior knowledge on alleles.

5.3.2 Output

*_calls_<method>.vcf.gz tab delimited text file containing meta-information lines, a header line, and data lines each containing information about a position in the genome.

5.3.3 Usage Example

# Simulate a BAM File
atlas simulate

# Call sites
atlas call --bam ATLAS_simulations.bam --fasta ATLAS_simulations.fasta