ljdursi / ga4gh_directory

A simple docker for firing up a GA4GH reference server from a directory of BAMs and VCFs

ga4gh_directory

This is a simple docker for starting up a GA4GH server based on a directory of BAMs and VCFs.

Take for example the GA4GH example data with indexes removed:

$ ls
GRCh37-subset.fa    HG00096.bam        HG00533.bam        HG00534.bam        chr1.vcf.gz        chr2.vcf.gz        chr3.vcf.gz

One can build the docker from this repo via docker build -t ga4gh_directory . or just pull the prebuilt image with docker pull quay.io/ljdursi/ga4gh_directory.

Above the data directory, you can then create the necessary indices and the registry.db file as so: (Note: the scripts currently assume that the mounted data directory is always /data)

$ docker run -v "${PWD}"/example_data:/data ga4gh_directory init /data/GRCh37-subset.fa /data/
[bgzip] No such file or directory: /data//*.vcf
Verifying ReferenceSet testref @ /data/GRCh37-subset.fa.gz
    Reading 1000 bases from 1
Verifying Dataset test
    Verifying ReadGroupSet HG00096 @ /data/HG00096.bam
        Read 10 alignments from SRR062634
        Read 10 alignments from SRR062635
        Read 10 alignments from SRR062641
    Verifying ReadGroupSet HG00533 @ /data/HG00533.bam
        Read 10 alignments from ERR020237
    Verifying ReadGroupSet HG00534 @ /data/HG00534.bam
        Read 10 alignments from ERR020238
    Verifying VariantSet testvars
        Read 10 variants from reference 1 @ /data/chr1.vcf.gz
        Read 10 variants from reference 3 @ /data/chr3.vcf.gz
        Read 10 variants from reference 2 @ /data/chr2.vcf.gz

$ ls example_data
GRCh37-subset.fa.gz     HG00096.bam      HG00533.bam.bai   chr1.vcf.gz        chr2.vcf.gz.tbi     registry.db
GRCh37-subset.fa.gz.fai HG00096.bam.bai  HG00534.bam       chr1.vcf.gz.tbi    chr3.vcf.gz
GRCh37-subset.fa.gz.gzi HG00533.bam      HG00534.bam.bai   chr2.vcf.gz        chr3.vcf.gz.tbi

From this same directory, we can now start up a ga4gh server, as well as serving the files themselves:

$ docker run -d -v "${PWD}"/example_data:/data -v /tmp:/var/log/apache2 -p 8000:80 ga4gh_directory serve
288fe71273c65d746e506c5a9ef06a29964535043adf3a4fa8550f05556a7c0a

$

and then:

$ curl http://localhost:8000/ga4gh/ --silent | grep GA4GH
        <title>GA4GH reference server 0.3.5</title>
        <h2>GA4GH reference server 0.3.5</h2>
    Welcome to the GA4GH reference server landing page! This page describes

$ wget -nv http://localhost:8000/chr3.vcf.gz
2017-01-25 00:42:15 URL:http://localhost:8000/chr3.vcf.gz [26967/26967] -> "chr3.vcf.gz" [1]