Tandem2XML issue with mutation searches

[dasaris]

Hi There,

I noticed an issue in Tandem2XML program. When the
program encounters a modification that is not in input parameters, it
issues an "unknown modification" warning and ignores it. However, this
often happens when people search for point mutations with X! Tandem. I
fixed the program such that it will write any unknown modifications to
the pepXML instead of ignoring them. Let me know if you are interested
in getting the fix into your source tree.

Thanks,
Surendra

[Brendan]

Hi Surendra,
The warning indicates that the correct modification specifications
were not written as <aminoacid_modification> tags to the
<search_summary> section of the pepXML. I guess just writing the
<mod_aminoacid_mass> tags in the <search_hits> might be an improvement
over just dropping them, but the real solution should be to figure out
why the <aminoacid_modification> tags aren't getting written.

Is there anything in the X! Tandem parameters than can be read as
indicating that these modifications are possible? Usually there is.

If you don't write internally consistent modifications, some tools
like CPAS (on which I was working when I wrote Tandem2XML) will have
difficulty consuming the resulting pepXML. I am sure Jimmy Eng or
Brian Pratt (included on this message) can help you connect with
people currently working on the TPP.

--Brendan

[dasaris]

Hi Brendan,

Currently, Tandem2XML scans the “potential modifications” tags of
tandem results file and writes them to the modification summary in the
pepXML. Point mutations would not be listed in “potential
modifications” tags of tandem results file. Only way you can find out
about them is to look for “point mutations” tag in the tandem XML. If
that tag is set to “yes”, then you have to pre-scan all peptide-
spectral matches (PSMs), build a table of point mutations seen in the
entire file, and write them to modification summary.

Another clean way of doing the same is to ignore the “potential
modifications” tag altogether. Pre-scan all PSMs for mods, and build a
table of seen mods, and write them to the modification summary. This
would option would work regardless of whether tandem is configured to
find mods or mutations. Both these options require a redesign of the
Tandem2XML source.

Meanwhile, I put in a stop gap fix that would write unknown mods to
pepXML instead of ignoring them. To me, ignoring them altogether is a
bigger problem than leaving them out of modification summary. Sure, it
produces inconsistent pepXMLs for programs that look at the
modification summary, but my feeling is that not a lot of programs
look at that summary.

Thanks,
Surendra

[Brendan]

Hi Surendra,
In light of the current state of the project, I agree with you. When
I put in the warning, I hoped to quickly fix any occurrences people
reported, and it did have an immediate positive effect. That warning
caused people to report issues, and I did fix a number of them.

The problem with ignoring is not knowing how to properly attribute the
modification. Was it variable or static?

There is no file that defines what point modifications are allowed?
It is really an arbitrary search where it calculates the modification
based solely on the data? If so, then maybe pepXML needs to be
modified to specify this type of modification well, as it does not
really fall into the categories covered by the current tags, I think.

Well, good luck. Always good to see people improving our tools.

--Brendan