UKB Meta-analysis

[yk-tanigawa]

We have a working version of meta-analysis summary statistics. When we finalize the summary statistics from each population, we should refresh the M-A as well.

[yk-tanigawa]

We pushed some jobs using a custom script.

Because we started this meta-analysis before completing the GWAS freeze, we have several versions of the M-A summary statistics in /oak/stanford/groups/mris/private_data/ukbb/24983/array-combined/metal/.

• 20200616: first attempt. There are 3760 traits covered in this analysis.
  • We have limited number of others and related summary statistics. (~50% of binary traits and ~99% of quantitative traits).
  • The summary statistics from 5 pops have issue #17
• 20200704: second attempt.
  • We started this freeze on the evening of 2020/7/4. Briefly, issue #17 is mostly fixed and others and related summary statistics are generated for almost all traits. See #21 for the remaining action items.
• 202007@@: we will generate another freeze when we finalize the summary statistics.

[yk-tanigawa]

version 20200704 is finished.

/oak/stanford/groups/mrivas/private_data/ukbb/24983/array-combined/metal/20200704/

3691 phenotypes.

[yk-tanigawa]

Version 20200717

ml load resbatch

cat ../../04_gwas/extras/202006-GWAS-finish/gwas-current-gz-wc.20200717-173655.annotated.tsv | egrep -v '^#' | cut -f1 | sort -u > 2_list_GBE_IDs.$(date +%Y%m%d-%H%M%S).lst

sbatch -p mrivas --qos=high_p --time=6:0:00 --mem=6000 --nodes=1 --cores=1 --job-name=metal --output=logs/metal.%A_%a.out --error=logs/metal.%A_%a.err --array=1-190 $parallel_sbatch_sh 1_metal.sh 2_list_GBE_IDs.20200717-204213.lst 20

Submitted batch job 4254457

[yk-tanigawa]

Metal is now finalized with 3794 traits /oak/stanford/groups/mrivas/ukbb24983/array-combined/metal/20200717

[yk-tanigawa]

Convert to plink format:

[ytanigaw@sh02-09n53 ~/repos/rivas-lab/ukbb-tools/18_metal/202006_metal]$ find /oak/stanford/groups/mrivas/ukbb24983/array-combined/metal/20200717 -name "*.gz" | sort > metal2plink.lst

[ytanigaw@sh02-09n53 ~/repos/rivas-lab/ukbb-tools/18_metal/202006_metal]$ sbatch -p mrivas --qos=high_p --time=1:0:00 --mem=6000 --nodes=1 --cores=1 --job-name=metal2plink --output=logs/metal2plink.%A_%a.out --error=logs/metal2plink.%A_%a.err --array=1-949 $parallel_sbatch_sh $(readlink -f ../metal_to_plink.sh) metal2plink.lst 4
Submitted batch job 5079692`