Correct Reference Genome For Cosmic

Question

[sfcarroll]

I am working with the COSMIC data extracts, specifically the UCSC Mutation Export, available at: ftp://ftp.sanger.ac.uk/pub/CGP/cosmic/data_export/Request_based_exports/UCSCMutExp_v66_250713.csv.gz

I am using the mutation locations in the file and then looking them up in the UCSC database. I noticed there is a FASTA archive hosted in the COSMIC ftp site. Should I be using this instead of the hg19.2bit file available from UCSC, or is the data the same? ftp://ftp.sanger.ac.uk/pub/CGP/cosmic/fasta.tgz

edit Related question: Applying SNP masking in primer creation

[Matt Shirley]

md5sum file1.fa file2.fa

If they contain the same info, the hashes will be identical.

[Hamish]

You might want to ask the COSMIC folks what they recommend, see http://cancer.sanger.ac.uk/cancergenome/projects/cosmic/contact

Answer 1

[Cyriac Kandoth]

You could use the raw COSMIC tab-delimited data instead of the request-based exports. It contains a mix of Build36 and Build37 loci. To parse through it, standardize everything to Build37, and fix several other problems, here is a script I recently posted to Github: https://github.com/ckandoth/parse-cosmic

And here is a list of standardized and annotated variants from Cosmic v64, with details on the caveats to parsing Cosmic data: https://www.synapse.org/#!Synapse:syn1855816