DEPICT code, instructions and an example
Switch branches/tags
Nothing to show
Clone or download

README.md

Dependencies

  • Mac OS X, or UNIX operating system (Microsoft Windows is not supported)
  • Java SE 6 (or higher)
  • Python version 2.7 (Python version 3 or higher is not supported)
  • PIP (used for install Python libraries)
    • sudo easy_install pip
  • Python intervaltree library
    • sudo pip install intervaltree
  • Pandas (version 0.15.2 or higher)
    • sudo pip install pandas
  • PLINK version 1.9 (August 1 release or newer)

DEPICT

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 DEPICT

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.

Test run DEPICT

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.

  1. Edit DEPICT/example/ldl_teslovich_nature2010.cfg
  • Point plink_executable to where PLINK executable (version 1.9 or higher) is on our system (e.g. /usr/bin/plink)
  1. Run DEPICT on the LDL summary statistics
  • E.g. ./src/python/depict.py example/ldl_teslovich_nature2010.cfg
  1. Investigate the results (see the Wiki for a description of the output format).
  • DEPICT loci ldl_teslovich_nature2010_loci.txt
  • Gene prioritization results ldl_teslovich_nature2010_geneprioritization.txt
  • Gene set enrichment results ldl_teslovich_nature2010_genesetenrichment.txt
  • Tissue enrichment results ldl_teslovich_nature2010_tissueenrichment.txt

Run DEPICT based on your GWAS

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.

  1. Make sure that you use hg19 genomic SNP positions
  2. Make an 'analysis folder' in which your trait-specific DEPICT analysis will be stored
  3. Copy the template config file from src/python/template.cfg to your analysis folder and give the config file a more meaningful name
  4. Edit your config file
  • Point analysis_path to your analysis folder. This is the directory to which output files will be written
  • Point gwas_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)
  • Specify the GWAS association p value cutoff (association_pvalue_cutoff). We recommend using 5e-8 or 1e-5
  • Specify the label, which DEPICT uses to name all output files (label_for_output_files)
  • Specify the name of the association p value column in your GWAS summary statistics file (pvalue_col_name)
  • Specify the name of the marker column (marker_col_name). Format: chr:pos, ie. '6:2321'. If this column does not exist chr_col and pos_col will be used, then leave if empty
  • Specify the name of the chromosome column (chr_col_name). Leave empty if the above marker_col_name is set
  • Specify the name of the position column (pos_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)
  • Specify the separator used in the GWAS summary statistics file (separator). Options are
    • tab
    • comma
    • semicolon
    • space
  • Point plink_executable to where PLINK 1.9 executable (August 1 release or newer) is on your system (e.g. /usr/bin/plink)
  • If you are using other genotype data than the data part of DEPICT then point 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
  1. Run DEPICT
  • <path to DEPICT>/src/python/depict.py <path to your config file>
  1. Investigate the results which have been written to your analysis folder. See the Wiki for details on the output format
  • Associated loci in file ending with _loci.txt
  • Gene prioritization results in file ending with _geneprioritization.txt
  • Gene set enrichment results in file ending with _genesetenrichment.txt
  • Tissue enrichment results in file ending with _tissueenrichment.txt

DEPICT at the Broad Institute

Run the LDL example

  1. Copy the example config file /cvar/jhlab/tp/depict/example/ldl_teslovich_nature2010.cfg to your working directory and change analysis_path to that directory
  2. Run DEPICT using qsub -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

Run DEPICT on own GWAS

  1. Follow the above steps 1-4
  2. Run DEPICT using
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.

Troubleshooting

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.

How to cite

Pers, Nature Communications 2015

1000 Genomes Project, because DEPICT makes extensively use of their data.

Data used in these examples

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.