High-Performance Computing at the NIH
GitHub YouTube @nih_hpc RSS Feed
LDSC on Biowulf

LDSC is a command line tool for estimating heritability and genetic correlation from GWAS summary statistics. LDSC also computes LD Scores.


The following is from https://github.com/bulik/ldsc#citation

Important Notes

One of the tests in this program's test suite is currentlly failing on Biowulf.
See https://github.com/bulik/ldsc/issues/64#issuecomment-319481128 for more information.

Interactive job
Interactive jobs should be used for debugging, graphics, or applications that cannot be run as batch jobs.

Allocate an interactive session and run the program. Sample session:

[user@biowulf]$ sinteractive
salloc.exe: Pending job allocation 46116226
salloc.exe: job 46116226 queued and waiting for resources
salloc.exe: job 46116226 has been allocated resources
salloc.exe: Granted job allocation 46116226
salloc.exe: Waiting for resource configuration
salloc.exe: Nodes cn3144 are ready for job

[user@cn3144 ~]$ module load ldsc
[+] Loading python 2.7 ...
[+] Loading ldsc, version 1.0.0-92-gcf1707e...
[user@cn3144 ~]$ wget https://data.broadinstitute.org/alkesgroup/LDSCORE/1kg_eur.tar.bz2
--2017-08-01 17:06:24--  https://data.broadinstitute.org/alkesgroup/LDSCORE/1kg_eur.tar.bz2
Resolving dtn01-e0...
Connecting to dtn01-e0||:3128... connected.
Proxy request sent, awaiting response... 200 OK
Length: 1134910 (1.1M) [application/x-bzip2]
Saving to: “1kg_eur.tar.bz2”

100%[==========================================================================================================>] 1,134,910    833K/s   in 1.3s    

2017-08-01 17:06:30 (833 KB/s) - “1kg_eur.tar.bz2” saved [1134910/1134910]

[user@cn3144 ~]$ tar -xf 1kg_eur.tar.bz2
[user@cn3144 ~]$ cd 1kg_eur/
[user@cn3144 ~]$ ldsc.py --bfile 22 --l2 --ld-wind-cm 1 --out 22
* LD Score Regression (LDSC)
* Version 1.0.0
* (C) 2014-2015 Brendan Bulik-Sullivan and Hilary Finucane
* Broad Institute of MIT and Harvard / MIT Department of Mathematics
* GNU General Public License v3
./ldsc.py \
--ld-wind-cm 1.0 \
--out 22 \
--bfile 22 \

Beginning analysis at Tue Aug  1 17:12:04 2017
Read list of 19156 SNPs from 22.bim
Read list of 379 individuals from 22.fam
Reading genotypes from 22.bed
After filtering, 19156 SNPs remain
Estimating LD Score.
Writing LD Scores for 19156 SNPs to 22.l2.ldscore.gz

Summary of LD Scores in 22.l2.ldscore.gz
         MAF        L2
mean  0.2323   18.5353
std   0.1453   16.1039
min   0.0013    0.0657
25%   0.1042    7.8392
50%   0.2243   13.4837
75%   0.3549   22.9722
max   0.5000  109.7163

MAF/LD Score Correlation Matrix
        MAF      L2
MAF  1.0000  0.2749
L2   0.2749  1.0000
Analysis finished at Tue Aug  1 17:12:08 2017
Total time elapsed: 3.31s
[user@cn3144 ~]$ exit
salloc.exe: Relinquishing job allocation 46116226
[user@biowulf ~]$

Batch job
Most jobs should be run as batch jobs.

Create a batch input file (e.g. ldsc.sh). For example:

module load ldsc

wget https://data.broadinstitute.org/alkesgroup/LDSCORE/1kg_eur.tar.bz2
tar -xf 1kg_eur.tar.bz2
cd 1kg_eur/
ldsc.py --bfile 22 --l2 --ld-wind-cm 1 --out 22

Submit this job using the Slurm sbatch command.

sbatch [--cpus-per-task=#] [--mem=#] ldsc.sh
Swarm of Jobs
A swarm of jobs is an easy way to submit set of independent commands requiring identical resources.

Create a swarmfile (e.g. ldsc.swarm). For example:

cd /data/$USER/ldsc/set1 && ldsc.py --bfile 22 --l2 --ld-wind-cm 1 --out 22
cd /data/$USER/ldsc/set2 && ldsc.py --bfile 22 --l2 --ld-wind-cm 1 --out 22
cd /data/$USER/ldsc/set3 && ldsc.py --bfile 22 --l2 --ld-wind-cm 1 --out 22
cd /data/$USER/ldsc/set4 && ldsc.py --bfile 22 --l2 --ld-wind-cm 1 --out 22

Submit this job using the swarm command.

swarm -f ldsc.swarm [-g #] [-t #] --module ldsc
-g # Number of Gigabytes of memory required for each process (1 line in the swarm command file)
-t # Number of threads/CPUs required for each process (1 line in the swarm command file).
--module ldsc Loads the TEMPLATE module for each subjob in the swarm