SDPR_admix

A statistical method to calculate PRS in admixed population. SDPR_admix is still in the early stage of testing. If you have encounbtered a problem, please contact zhou1633@purdue.edu.

Installation

To install SDPR, you need to first download the repo:

git clone https://github.com/eldronzhou/SDPR_admix.git

You can then compile SDPR by running make. If there is run time error that the shared library "libgsl.so.0" not found, you can fix it by typing

export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:path_to_SDPR_admix_dir/gsl/lib

Quick start

SDPR_admix can be run from the command line. To see the full list of options, please type

./SDPR_admix -h

Below are the required options.

• vcf (required): Path to the phased genotype file in the gzipped vcf format. Not that currently missing genotypes are not supported. If you use Eagle2 and Rfmix2, there shouldn't be missing genotype in the default setting.
• msp (required): Path to path to the directory containing RFMix2 solved local ancestry files.
• pheno (required): path to the phenotype file. The phenotype will be read from the 3rd column of the specified space- or tab-delimited file. There is no header and NA value can be included.
• covar (required): path to the covariate file. Covariates will be reading from the first column. There is no header for the covariate file.
• out (required): Path to the output file containing estimated effect sizes.
• rho (required): Trans-ethnic genetic correlation output by Admix-kit between 0 and 1. Default is 0.8.

Running SDPR_admix

./SDPR_admix -vcf test/chr22_train.vcf.gz -msp test/chr22_train.msp.tsv -pheno test/train.pheno -covar test/covar.txt -out test/res.txt

The output file has the following format:

1       833068  rs12562034      G       A       0       0
1       843942  rs4040617       A       G       0       0
...

where the columns are chromsome, position, variant ID, effect allele, non-effect allele, effect sizes corrsponding to population 0 in RFMix2 file and effect sizes corrsponding to population 1 in the RFMix2 file.

Derive the PRS

The following command can be used to calculate ancestry-aware PRS for the admixed population.

./score -vcf test/chr22_train.vcf.gz -msp test/chr22_train.msp.tsv -score test/res.txt -out test/test.profile

If you have bgzipped vcf and RFMix2 local ancestry files for chr1-22 with the name test/chr[1-22]_train.vcf.gz, then you can use the following command for iterative calculation over all chromsomes:

./score -vcf test/chr#_train.vcf.gz -msp test/chr#_train.msp.tsv -score test/res.txt -out test/test.profile