GA4GH Regions of Interest (FASTA Files)

Regions:

• BRCA1
• BRCA2
• SMA
• KIR
• MHC

Overview

• Quickstart: How to download the FASTA files
• Extract Sequences Yourself: Describes how to independently download the regions of interest
• About: Meta-information about this project

Quickstart

I created a collection of the FASTA files for the above regions of interest as defined by NCBI Gene and fetched from NCBI Nucleotide (the list of NCBI Gene IDs for each region was compiled by hand).

The FASTA files are in the FASTA directory and total ~8 MB. They are named GeneName-NCBIGeneID.fa for searching convenience.

FASTA
    HLA
	    (28 files)
    KIR
	    (17 files)
    SMA-6606.fa
    BRCA2-675.fa
    BRCA1-672.fa

Preview the collection contents

• https://workbench.qr1hi.arvadosapi.com/collections/qr1hi-4zz18-7zk4muy5grnaqpv#
• (deprecated-but-functional sharing link) https://workbench.qr1hi.arvadosapi.com/collections/download/qr1hi-4zz18-7zk4muy5grnaqpv/2sf3n5y9nuqhi815xioxa2ychioyqts7i5ysxg5we8fdl988az/

Besides the FASTA directory, there are two other directories, NCBIGeneXML and GeneRefSeqCoordinates, which contain the intermediate output used to fetch the FASTA files. In total, the collection size is ~65 MB.

Download the entire collection contents

A. if you have access to an Arvados VM

arv get qr1hi-4zz18-7zk4muy5grnaqpv . #note trailing dot!

B. otherwise, they are publicly available

wget --mirror --no-parent --no-host --cut-dirs=3 https://workbench.qr1hi.arvadosapi.com/collections/download/qr1hi-4zz18-7zk4muy5grnaqpv/2sf3n5y9nuqhi815xioxa2ychioyqts7i5ysxg5we8fdl988az/

Extract Sequences Yourself

In short:

$ git clone <this repository>
$ sh ./geneID_to_refseq_regions.sh ids.csv

Note that the Entrez Direct tools must be available on your PATH.

This describes the process I used to create these FASTA files. First, compile a list of GeneIDs for each region. I did this by hand in a google doc and the results are at in this repository: ids.csv.

BRCA1,672
BRCA2,675
[...]
V,352962

Note: For the HLA genes, I used the IMGT list of gene names, found at ftp://ftp.ebi.ac.uk/pub/databases/ipd/imgt/hla/fasta/ . For the KIR genes, I used NCBI Gene query to list all the KIR genes. http://www.ncbi.nlm.nih.gov/gene and "Homo sapiens"[porgn] AND KIR

Then, modify the geneID_to_refseq_regions.sh file to run as many times as there are genes / lines in the ids.csv file.

$ vi geneID_to_refseq_regions.sh
	    for run in {1..48} #change this!

Then, run the script against the CSV file of genes. Make sure that you have xsltproc installed if you do not already, and that transform.xsl is in your root directory.

$ sudo apt-get install xsltproc
$ sh ./geneID_to_refseq_regions.sh ids.csv

All files will be output to the root directory. Sort by filetype and move into separate folders as you wish.

About

Version

• v0.2 -- (08 Dec 2014) KIR, HLA gene list compiled. All five regions fetched from NCBI. Credit for transform.xsl goes to Pierre Lindenbaum, 06 Dec 2014, https://www.biostars.org/p/122680/#122689 .
• v0.1 -- (02 Dec 2014) BRCA1, BRCA2, SMA genes on hg19 and hg38 FASTA files created.

Notes

• v0.2 -- The previous version can be found at https://gist.github.com/nouyang-curoverse/b17a3820d2ccf36f3e47
• v0.1 -- The python exactsearch.py program is documented in the source code at https://gist.github.com/nouyang-curoverse/b17a3820d2ccf36f3e47

License

Public Domain

Author

Nancy Ouyang (nancy@curoverse.com)