Mutation file format docs do not mention correct column for Amino_Acid_Change about cbioportal

[ecerami]

Actually, it seems that the only column where you can specify amino acid change must now be specified as: HGVSp_Short. Is this on purpose, or maybe just temporary? @schultzn @n1zea144 @jjgao

[onursumer]

This is on purpose. HGVSp_Short represents the protein change value coming from the VEP annotator.

We also have another column (Amino_Acid_Change) for custom annotations, but this value has not made its way to the front-end yet.

[ecerami]

but, what about others who don't have VEP annotated data?  (this is true of most other installs).  this also doesn't match the format details on the wiki.  can we consider a more generic solution?

Sent from mobile phone.  Please excuse typos.

-------- Original message --------
From: onursumer [email protected]
Date:05/13/2015 3:18 PM (GMT-05:00)
To: cBioPortal/cbioportal [email protected]
Cc: ecerami [email protected]
Subject: Re: [cbioportal] Mutation file format docs do not mention correct
column for Amino_Acid_Change (#17)

This is on purpose. HGVSp_Short represents the protein change value coming from the VEP annotator.

We also have another column (Amino_Acid_Change) for custom annotations, but this value has not made its way to the front-end yet.

—
Reply to this email directly or view it on GitHub.

[n1zea144]

Perhaps we should support an annotation property in the meta file for mutations? Based on this property the import code does the right thing?
On May 13, 2015, at 5:27 PM, ecerami [email protected] wrote:

but, what about others who don't have VEP annotated data? (this is true of most other installs). this also doesn't match the format details on the wiki. can we consider a more generic solution?

Sent from mobile phone. Please excuse typos.

-------- Original message --------
From: onursumer [email protected]
Date:05/13/2015 3:18 PM (GMT-05:00)
To: cBioPortal/cbioportal [email protected]
Cc: ecerami [email protected]
Subject: Re: [cbioportal] Mutation file format docs do not mention correct column for Amino_Acid_Change (#17)

This is on purpose. HGVSp_Short represents the protein change value coming from the VEP annotator.

We also have another column (Amino_Acid_Change) for custom annotations, but this value has not made its way to the front-end yet.

—
Reply to this email directly or view it on GitHub.

—
Reply to this email directly or view it on GitHub.

[ecerami]

A while back (back in the old days!), I remember we had this same problem in TCGA where each sequencing center labeled the amino acid change differently, e.g. AMINO_ACID_CHANGE, AA_CHANGE, PROTEIN_CHANGE, etc. The MAF reader had the "smarts" to figure out which columns were specified, and pick out the right amino acid change column. If we did this, then the code would always work, regardless of MAF variations, and no need to configure anything. Thoughts?

[schultzn]

I thought are still supporting Protein_Change in the absence of VEP annotations?

On May 14, 2015, at 11:25 AM, ecerami [email protected] wrote:

A while back (back in the old days!), I remember we had this same problem in TCGA where each sequencing center labeled the amino acid change differently, e.g. AMINO_ACID_CHANGE, AA_CHANGE, PROTEIN_CHANGE, etc. The MAF reader had the "smarts" to figure out which columns were specified, and pick out the right amino acid change column. If we did this, then the code would always work, regardless of MAF variations, and no need to configure anything. Thoughts?

—
Reply to this email directly or view it on GitHub.

[ecerami]

@schultzn: As far as I can tell, only HGVSp_Short is recognized. Hence, if you have a MAF with Protein_Change or AA_Change, mutations just appear as "MUTATED", but have no amino acid change details.

[jjgao]

I think we can fall back on Amino_Acid_Change if HGVSp_Short does not exist or is empty.

[ecerami]

Ah, that's true. AMINO_ACID_CHANGE is in the code. But, if it's specified, it is stored in a separate field called aminoAcidChange, and only proteinChange is actually stored in the database.

I am looking at this: https://github.com/cBioPortal/cbioportal/blob/master/core/src/main/java/org/mskcc/cbio/maf/MafUtil.java

record.setProteinChange(TabDelimitedFileUtil.getPartString(getColumnIndex(PROTEIN_CHANGE), parts));

and

public static final String PROTEIN_CHANGE = "HGVSp_Short";

And in ExtendedMutationUtil:

    String proteinChange = record.getProteinChange();

    // if protein change is not valid, then use the string "MUTATED"
    if (!isValidProteinChange(proteinChange))
    {
        proteinChange = "MUTATED";
    }

[onursumer]

I think if AMINO_ACID_CHANGE column exists in the MAF but not HGVSp_Short column, then we should set the protein change value to AMINO_ACID_CHANGE column value. Do we also need to consider other custom column names to get protein change value, or does AMINO_ACID_CHANGE suffice?

[onursumer]

Now importing Amino_Acid_Change column with the commit 9a2f1cd

[ecerami]

Thanks, @onursumer! I will update the wiki later today, and close out...

[ecerami]

Hi @onursumer. I am trying to find this bug fix. Do you know which branch you committed to? Tx.

[onursumer]

Somehow it created another tree in master branch, I think I was out of
sync. Just merged and pushed it to the master branch. Let me know if it
works now.

On Mon, May 18, 2015 at 4:54 PM, ecerami [email protected] wrote:

Hi @onursumer https://github.com/onursumer. I am trying to find this
bug fix. Do you know which branch you committed to? Tx.

—
Reply to this email directly or view it on GitHub
#17 (comment)
.

[ecerami]

Hurray! this now works. I have also updated the wiki doc to explain this: https://github.com/cBioPortal/cbioportal/wiki/File-Formats#mutation-data. Closing...