phase_common [formerly known as SHAPEIT4]

Table of contents

1. Description
2. Usage1: phasing unrelated samples
3. Usage2: phasing related samples
4. Usage3: phasing chromosome X data
5. Usage4: phasing using a reference panel
6. Usage5: phasing using a scaffold
7. Command line options
   1. Basic options
   2. Input files
   3. Filter parameters
   4. MCMC parameters [Expert]
   5. PBWT parameters [Expert]
   6. HMM parameters [Expert]
   7. Output files

Description

Tool to phase common variant sites, typically SNP array data (i.e. prephasing). This tool is also used as the first step for phasing large WES/WGS datasets. This is an improved version of SHAPEIT4 and should serve as a replacement.

Usage1: phasing unrelated samples

Go in the test folder and run:

phase_common --input array/target.unrelated.bcf --region 1 --map info/chr1.gmap.gz --output target.phased.bcf --thread 8

The program phases data in the input (unphased) file (--input array/target.unrelated.bcf):

• using 8 threads (--thread 8)
• data located on chromosome 1 (--region 1)
• using a specific genetic map (--input info/chr1.gmap.gz)
• saving the phased haplotypes in output file (--output target.phased.bcf).

If no genetic map is specified, a recombination rate of 1 cM/Mb is assumed by default.

Usage2: phasing related samples

Go in the test folder and run:

phase_common --input array/target.family.bcf --pedigree info/target.family.fam --region 1 --map info/chr1.gmap.gz --output target.phased.bcf --thread 8

The family relationships (parent-offspring relationships, duos/trios) are specified in the file info/target.family.fam. This file contains one line per sample having parent(s) in the dataset and three columns (kidID fatherID and motherID), separated by TABs for spaces. Use NAs for unknown parents (in the case of duos). In output file, the first offspring haplotype is transmitted by the father, the second by the mother.

This tools uses family data to fix the phase of offspring heterozygous genotypes when possible, that is when:

• There is no Mendel inconsistency,
• At least one parent is homozygous. In other words, it builds a scaffold of haplotypes for offsprings from the parental genomes.

Usage3: phasing chromosome X data

Go in the test folder and run:

phase_common --input array/target.haploid.bcf --haploids info/target.haploid.txt --region 1 --map info/chr1.gmap.gz --output target.phased.bcf --thread 8

The list of samples being haploid (e.g. males for chromosome X) is specified in the file info/target.haploid.fam. This file contains one line per sample.

For convenience, haploid samples are encoded either internally or in input/output as fully homozgote diploid samples. In practice, this tool sets to missing any heterozygous genotype in input and re-impute them as REF/REF or ALT/ALT. For all samples not listed in the haploid file, the tool proceeds with standard phasing.

Usage4: phasing using a reference panel

Go in the test folder and run:

phase_common --input array/target.unrelated.bcf --reference array/reference.bcf --region 1 --map info/chr1.gmap.gz --output target.phased.bcf --thread 8

The tool uses the haplotypes specified in the BCF/VCF file array/reference.bcf as a reference panel to phase all the data in input file. Any input site not in the reference panel is not considered in the phasing.

Usage5: phasing using a scaffold

Go in the test folder and run:

phase_common --input array/target.unrelated.bcf --scaffold array/target.scaffold.bcf --region 1 --map info/chr1.gmap.gz --output target.phased.bcf --thread 8

The tool uses the haplotypes specified in the BCF/VCF file array/target.scaffold.bcf as a scaffold to phase all the data in input file. The scaffold can contain data for a subset of samples at a subset of variant sites. Data in the scaffold needs to be non-missing and phased.

Command line options

Basic options

Input files

Filter parameters

MCMC parameters [Expert]

PBWT parameters [Expert]

HMM parameters [Expert]

Output files