ibp_pgs_pipelines

This is a nextflow pipeline designed to calculate polygenic scores at IBP

The pipeline can be cloned along with all its dependencies using the following command:

git clone --recursive https://github.com/vaqm2/ibp_pgs_pipelines.git

The pipeline can be run as follows to display the parameter options and help message:

nextflow run main.nf --help

The inputs required

• A reference GWAS in VCF format, indexed using tabix. These can be obtained from the output of the IBP sum stats cleaning pipeline: https://github.com/BioPsyk/cleansumstats
  Please make sure that the VCF has standard errors in the format column as the older summary stats might be missing these. The code ignores SNPs without standard errors.
  The file to use from the output directory is labeled: cleaned_GRCh37_full_altbased.vcf.gz

• A json file containing paths to PLINK format .bed/.bim/.fam files. These can be obtained for iPSYCH datasets on iPSYCH-GDK and a 1000 Genomes test dataset for GDK open from the JSON/ folder that comes with the pipeline

• Linkage disequilibrium matrices for the SNPs used to calculate PGS These are hardcoded in the nextflow.config file for iPSYCH-GDK and gdk_open.config for GDK-open

• A phenotype file with the columns: FID IID [Phenotype_Name]

• A covariate file with the columns: FID IID Age Sex PC1 PC2... This is hardcoded in the nextflow.config file for iPSYCH and gdk_open.config for GDK-open.

• The pipeline requires the R packages ggplot2, dplyr, fmsb, data.table, argparser

• The pipeline requires the python packages scipy, h5py, cyvcf2

The file rquirements.txt produced by conda can be used to reproduce the developer environment to run this pipeline with the following command: conda create --name <env> --file requirements.txt

The outputs produced

• A file containing PGS from different methods. For example, if your trait is named BMI, the score file will be named: iPSYCH2012_All_Imputed_2021_QCed_BMI_Scores.txt with the columns:

    IID [ Sample identifer ]
    PT_5E8 [ PRSice pruning and thresholding score with SNPs P < 5E-8 ]
    PT_1E6 [ PRSice pruning and thresholding score with SNPs P < 1E-6 ]
    PT_0.05 [ PRSice pruning and thresholding score with SNPs P < 0.05 ]
    PT_1 [ PRSice pruning and thresholding score with all SNPs ]
    sBayesR_UKBB_2.8M  [ sBayesR scores using UKBB as reference LD and 2.8M high quality SNPs ]
    sBayesR_UKBB_HM3 [ sBayesR scores using UKBB as reference LD and HapMap3 SNPs ]
    PRSCS_UKBB_HM3 [ PRS-CS scores using UKBB as reference LD and HapMap3 SNPs ]
    PRSCS_1KG_HM3` [ PRS-CS scores using 1000 Genomes as reference LD and HapMap3 SNPs ]

• A file containing the variance explained in phenotype by each score with columns

    Method
    r2 [ Nagelkerke R2 ]
    r2_L [ Liability adjusted R2 for case-control outcomes ]
    P

• A bar plot for variance explained

NOTE

When launching the pipeline as an sbatch script, please be mindful of requesting long enough wall-time for other more computationally expensive jobs, such as calculating SNP posteriors to run and finish. This depends on how busy the queue is, but usually around 48h is long enough on GenomeDK.