sudo easy_install pip
sudo pip install intervaltree
sudo pip install pandas
The following description explains how to download DEPICT, test run it on example files and how to run it on your GWAS summary statistics.
Download the compressed DEPICT version 1 rel194 files and unzip the archive to where you would like the DEPICT tool to live on your system. Note that you when using DEPICT can write your analysis files to a different folder. Be sure to that you meet all the dependencies described above. If you run DEPICT at the Broad Institute, see below section.
The following steps outline how to test run DEPICT on LDL cholesterol GWAS summary statistics from Teslovich, Nature 2010. This example is available in both the 1000 Genomes Project pilot phase DEPICT version and the 1000 Genomes Project phase 3 DEPICT version.
DEPICT/example/ldl_teslovich_nature2010.cfg
plink_executable
to where PLINK executable (version 1.9 or higher) is on our system (e.g. /usr/bin/plink
)./src/python/depict.py example/ldl_teslovich_nature2010.cfg
ldl_teslovich_nature2010_loci.txt
ldl_teslovich_nature2010_geneprioritization.txt
ldl_teslovich_nature2010_genesetenrichment.txt
ldl_teslovich_nature2010_tissueenrichment.txt
The following steps allow you to run DEPICT on your GWAS summary statistics. We advice you to run the above LDL cholesterol example before this point to make sure that you meet all the necessary dependencies to run DEPICT.
src/python/template.cfg
to your analysis folder and give the config file a more meaningful nameanalysis_path
to your analysis folder. This is the directory to which output files will be writtengwas_summary_statistics_file
to your GWAS summary statistics file. This file can be either in plain text or gzip format (i.e. having the .gz extension)association_pvalue_cutoff
). We recommend using 5e-8
or 1e-5
label_for_output_files
)pvalue_col_name
)marker_col_name
). Format: chr_col_name
). Leave empty if the above marker_col_name
is setpos_col_name
). Leave empty if the above marker_col_name
is set. Please make sure that your SNP positions used human genome build GRCh37 (hg19)separator
). Options aretab
comma
semicolon
space
plink_executable
to where PLINK 1.9 executable (August 1 release or newer) is on your system (e.g. /usr/bin/plink
)genotype_data_plink_prefix
to where your PLINK binary format 1000 Genomes Project genotype files are on your system. Specify the entire path of the filenames except the extension<path to DEPICT>/src/python/depict.py <path to your config file>
_loci.txt
_geneprioritization.txt
_genesetenrichment.txt
_tissueenrichment.txt
/cvar/jhlab/tp/depict/example/ldl_teslovich_nature2010.cfg
to your working directory and change analysis_path
to that directoryqsub -e err -o out -cwd -l h_vmem=12g /cvar/jhlab/tp/depict/src/python/broad_run.sh python /cvar/jhlab/tp/depict/src/python/depict.py <your modified config file>.cfg
use UGER
qsub -e err -o out -cwd -l m_mem_free=2.5g -pe smp 6 /cvar/jhlab/tp/depict/src/python/broad_run.sh python /cvar/jhlab/tp/DEPICT/src/python/depict.py <your modified config file>.cfg
Be aware that DEPICT needs at least needs 14GB memory when if modify the memory used per slot/thread.
Please send the log file (ending with _log.txt
) with a brief description of the problem to Tune H Pers (tunepers@broadinstitute.org).
The overall version of DEPICT follows the DEPICT publications. The current version is v1
from Pers, Nature Communications, 2015 and the release follows the number of commits of the DEPICT git repository (git log --pretty=format:'' | wc -l
). The latest 1000 Genomes Project pilot phase DEPICT version is rel138
, the latest 1000 Genomes Project phase 3 version is rel137
.
Pers, Nature Communications 2015
1000 Genomes Project, because DEPICT makes extensively use of their data.
LDL GWAS summary statistics from Teslovich, Nature 2010 are used as input in this example. We included all SNPs with P < 5e-8 and manually added chromosome and position columns (hg19/GRCh37).
1000 Genomes Consortium pilot release and phase 3 release data are used in DEPICT. Please remember to cite their paper in case you use our tool.