Data preparation

Goal

Clean and align raw genotype data into a VCF that the phasing software can consume — proper chromosome naming, biallelic filtering, and a genetic map consistent with sample-id conventions. This is the first stage of the four-stage pipeline.

Pipeline position

Inputs: raw VCF from a variant caller, or PLINK .bed/.bim/.fam, or genotype CSV + genetic-map text file.
Outputs: a single, biallelic, unphased VCF ready for Beagle.
Tool: plink (for format conversion) or convert-to-vcf (if starting from CSV); no maspipeline binary required for the VCF-only path.
Previous stage: — (this is the entry point)
Next stage: Phasing

Raw genotype data is expected in a .VCF format, as it contains genetic map positions, reference/alternative alleles, and unphased genotype calls (e.g., 0/1), which serve as the essential input for the subsequent phasing step. However, if the raw genotype data is in PLINK formats (provided as .bed, .bim, .fam), recoding to .VCF format can be done by PLINK as well, for example:

plink --bfile GENO_DATA --chr-set 95 no-xy --recode vcf --out GENO_DATA

This will generate unphased genotype data in a .VCF file, for example:

##fileformat=VCFv4.2
##fileDate=20250503
##source=PLINKv1.9
##contig=<ID=1,length=80480704>
##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">
#CHROM  POS ID  REF ALT QUAL    FILTER  INFO    FORMAT  Ind_1   Ind_2
1   16977   snp1    C   A   .   .   PR  GT  0/0 0/1
1   33723   snp2    G   T   .   .   PR  GT  0/0 0/0

Alternatively, if the genotype data is in form of .CSV:

ID,snp_1,snp_2,snp_3,...snp_n
ind_1,0,1,2,...
ind_2,1,2,0,...
ind_3,2,0,1,...

Along with a genetic map (in form .txt):

SNPID   Chr Position
snp_1   1   0
snp_2   1   16746
snp_3   1   217334

We provide convert-to-vcf to convert such genotype data to a VCF file. This tool can be run as follows:

./convert-to-vcf \
      -i genotypes.csv \
      -m map.txt \
      -o genotypes.vcf \
      --split-by-chr \
      --parallel \
      --ncores 4 \
      --merge-after-split

Note that the conversion is done by splitting the genotype files based on the number of chromosomes processed in parallel, then merging them into a single VCF file.

What’s next?

→ Continue to Phasing — resolve diploid genotypes into haploid maternal/paternal strands with Beagle.

← Back to overview.