I am a computer scientist studying pattern recognition, and I am hoping to do some supervised learning on colon cancer. Unfortunately, I'm having a heck of a time finding DNA data in the following format.

Benign DNA (Adenoma?)     Malignant DNA (Carcinoma?)
A                         A
A                         A
T                         G
G                         G
C                         C

So I just need two columns (don't care if it goes down for 3 billion records/ amino acids): one with the good DNA and one with the bad.

Is there anywhere I can find data in that specific format?

  • 1
    $\begingroup$ Did you checkout dcc.icgc.org? I think unlikely that someone will have the data formatted precisely the way you described, but you should be able to reformat it to suit your analysis pipeline. What you can download from there is a VCF file which only lists the positions in the tumour where it differs from the Normal so if you really want the entire genome file you would need to start with the human reference sequence, and change the sites to match the annotations in the VCF file. N.B. the will not have a comparison between a benign and a malignant. just tumour vs. normal $\endgroup$ – mdperry Dec 1 '15 at 18:33

You may try the data for "Colon Adenocarcinoma" made by the TCGA project: http://gdac.broadinstitute.org/runs/analyses__2015_04_02/reports/cancer/COAD/

The file with mutations called by the tumor against a matched normal, http://gdac.broadinstitute.org/runs/analyses__2015_04_02/reports/cancer/COAD/MutSigNozzleReport2.0/COAD-TP.final_analysis_set.maf

What you need are column 11, 12 and 13. Column 11 is the reference allele (your so-call good DNA). For each row, the alternative allele (bad DNA) is column 12 if column 12 does not equal column 11, otherwise it is column 13. -- It would be a piece of cake to make by coding :)

In case you are interested, column 16 and 17 are the tumor and normal sample IDs.

  • $\begingroup$ THANK YOU!!! This is exactly what I need, and thank you for making the field references easy. I'm using R to extract the .maf format, so I very much appreciate how much easier your answer makes the data understanding. Thank you as well for the insight into columns 16 and 17. Really, very helpful, thank you. $\endgroup$ – Jonathan Charlton Dec 1 '15 at 19:53
  • 1
    $\begingroup$ You're most welcome! Here is more information about the .maf file, biostars.org/p/69222 , though it might be a bit biology intensive. $\endgroup$ – xb. Dec 1 '15 at 19:58
  • $\begingroup$ One quick question about that data: a couple of the alleles (records for the alleles) have more than one letter. Do you know what that is about? Just want to know if I should record it as an A, T, G, or a C. Thank you so much again. $\endgroup$ – Jonathan Charlton Dec 1 '15 at 21:49
  • 1
    $\begingroup$ Those are indels (insertions or deletions) that involve multiple nucleotides. The rest are so-called SNVs, single nucleotide variations. Splitting and merging over indels/SNVs is a common practice in data analysis. $\endgroup$ – xb. Dec 2 '15 at 13:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.