wanglab/spark-inferno repository overview

⁠Spark-INFERNO

Spark-INFERNO is a scalable high-throughput pipeline for inferring the molecular mechanisms of non-coding genetic variants

⁠Spark-INFERNO source repository

https://bitbucket.org/wanglab-upenn/sparkinferno⁠

⁠Using Spark-INFERNO docker image

1. Install annotation data repository

   docker run -v [target_annot_dir]:[target_annot_dir] wanglab/spark-inferno /opt/install_annot.sh [target_annot_dir] [annotation_metafile_URL]

   For example, to install annotation data repository for GRCh37/hg19 genome build (https://tf.lisanwanglab.org/GADB/metadata/metadata.latest.hg19.template⁠) into /mnt/data/annotation_data_hg19 directory:

   docker run -v /mnt/data/annotation_data_hg19:/mnt/data/annotation_data_hg19 wanglab/spark-inferno /opt/install_annot.sh /mnt/data/annotation_data_hg19 https://tf.lisanwanglab.org/GADB/metadata/metadata.latest.hg19.template
   

   Copy

   GRCh38/hg38 annotation data repository could be similarly installed using hg38 metadata URL: https://tf.lisanwanglab.org/GADB/metadata/metadata.latest.hg38.template⁠

2. Run Spark-INFERNO pipeline (spark_inferno.py)

   2.1. Get Bash session inside docker:

    docker run -it -v /mnt/data/annotation_data_hg19:/mnt/data/annotation_data_hg19 -v /mnt/data/spark_inferno:/mnt/data/spark_inferno wanglab/spark-inferno /bin/bash
   

   Copy

   NOTE: A minimum of two host directories need to be provided for docker container:

   1. annotation data directory (/mnt/data/annotation_data_hg19 in the example command above)
   2. working directory (/mnt/data/spark_inferno in the example command above). Working directory could contain necessary files for running Spark-INFERNO analysis e.g., input files (GWAS summary statistics files, top SNPs files, etc), configuration files and output directories with Spark-INFERNO analysis results.

   2.2. Then run Spark-INFERNO: spark_inferno.py --help

Detailed documentation on running Spark-INFERNO and usage examples are available at https://bitbucket.org/wanglab-upenn/sparkinferno⁠ (README.md)