SHAPEIT using VCF unphased genotype input

I can get SHAPEIT to work with the default Plink PED/MAP format input files, but not with a VCF as input.

As an example, here I use the demo data that comes with SHAPEIT, which runs well.

DEMO=/Users/michaelflower/bin/shapeit.v2.904.3.10.0-693.11.6.el7.x86_64/example

shapeit -B $DEMO/gwas.bed $DEMO/gwas.bim $DEMO/gwas.fam \
-M $DEMO/genetic_map.txt \
-O "$DIR"/shapeit/gwas.phased

However, when I try and use the VCF file they provide it errors.

gunzip "$DIR"/shapeit/demo/gwas.vcf.gz

Initially when I run with --input-vcf, like this:

shapeit --input-vcf $DIR/shapeit/demo/gwas.vcf \
M $DEMO/genetic_map.txt \
-O "$DIR"/shapeit/gwas.phased

I get the error:

Phaser mode : unrecognised option '--input-vcf'

And when I try with the abbreviated format:

shapeit -V $DIR/shapeit/demo/gwas.vcf \
M $DEMO/genetic_map.txt \
-O "$DIR"/shapeit/gwas.phased

I just get this printed in the terminal, but no output files are produced in the output directory.

Segmented HAPlotype Estimation & Imputation Tool
  * Authors : Olivier DELANEAU, Jean-François ZAGURY & Jonathan MARCHINI
  * Contact : olivier.delaneau@gmail.com
  * Webpage : http:://www.shapeit.fr
  * Version : v2.r648

I'd be very grateful for a little help to get this working, thanks

vcf shapeit

• 2,927 views • link

written 4.7 years ago by michael.flower.14 •

You should strongly consider using shapeit4 - shapeit2 is about 10 years old now.

• 0 views • link

written 4.7 years ago by 4galaxy77

Thanks, I'm trying to install shapeit4 with conda, but am getting:

PackagesNotFoundError: The following packages are not available from current channels:
  - shapeit4

• 0 views • link

written 4.7 years ago by michael.flower.14 •

1 answer

I managed to solve this by converting the VCF to plink format

#=================================================================
# Convert VCF to plink format
#=================================================================
# https://www.biostars.org/p/207388/
# https://www.cog-genomics.org/plink2/data#recode

# Install plink
#conda install -c bioconda plink
conda create -n plink -c conda-forge -c bioconda plink

# Enter plink environment
conda activate plink

# Set VCF shortcut
VCF="$DIR"/wgs/130iPSC_061118.snp.vcf.gz

# Convert to plink binary format (bed, bim, fam)
plink --vcf "$VCF" --out "$DIR"/plink/$PREFIX

# Convert to plink ped format (ped, map)
plink --vcf "$VCF" --recode --out "$DIR"/plink/ped/$PREFIX



# For shapeit each "chromosome" needs to have its own input file
# ShapeIT can just phase one chromosome at a time
# https://bioinformatics.stackexchange.com/questions/2883/problem-of-ordering-in-physical-positions-phasing-snps-with-shapeit

for chr in $(seq 1 22) ; do plink --file "$DIR"/plink/ped/$PREFIX --chr $chr --recode --out "$DIR"/plink/ped/$PREFIX"_chr"$chr ; done

# Exit plink environment
conda deactivate

• 0 views • link

written 4.7 years ago by michael.flower.14 •