MaxQuant can have long-format modification notations; does not work in current implementation

[NoeGuilloy]

Hi,

I am running ms2pip through ms2resocre, and i get that following Index error.

Merging results multiprocessing.pool.RemoteTraceback: """ Traceback (most recent call last): File "/home/noeguill/anaconda3/lib/python3.7/multiprocessing/pool.py", line 121, in worker result = (True, func(*args, **kwds)) File "/home/noeguill/anaconda3/lib/python3.7/site-packages/ms2pip/ms2pipC.py", line 316, in process_spectra modpeptide = apply_mods(peptide, mods, PTMmap) File "/home/noeguill/anaconda3/lib/python3.7/site-packages/ms2pip/ms2pipC.py", line 631, in apply_mods modpeptide[pos] = mod IndexError: index 23 is out of bounds for axis 0 with size 19 """

Any ideas ?

Thanks a lot, Noé

[NoeGuilloy]

Hi, It looks like the error is coming from ms2rescore. I am using the ms2rescore maxquant pipeline.

I tried to run it again with a different sample and I get the same error but this time with Index 43 is out of bound.

With a grep ‘43|’ on the msms.peprec I saw that strange line as the C is not in 43 and there is only 21 AA in that sequence. AML_P07_G1_F1.15742.15742 SMMQDREDQSILCTGESGAGK 43|Carbamidomethyl 3 ENSP00000216181.10 66.874 43.867 1

I looked at the same sequence in the msms.txt file SM(Oxidation (M))M(Oxidation (M))QDREDQSILCTGESGAGK

Here if you count the variable mods the C become in position 43 as in the peprec file. It seems that if there is a fixed modification after a variable modification, then the fixed one is not as the right place in the prerec.

Is there something I do wrong? or do you know any fixes?

Thanks, Noé

[RalfG]

Hi Noé,

Thanks for using MS²PIP and MS²ReScore! It could be that this error is the result in a bug which will be fixed in an upcoming release. The bug was introduced in v3.6.3, so could you try v3.6.2 instead and check whether this fixes the issue?

Thanks! Ralf

[NoeGuilloy]

Hi,

I couldn't use v3.6.2 as it wasn't working with ms2rescore. However it seem that I the version of MaxQuant (1.6.14) i use don't have the same 2 letters code for PTMs. I tried to change it in the config-file of ms2rescore didn't work. Changing the annotation by replacing the new code by the 2 letters code in the msms.txt worked.

Also if I wanted to use only the Ms2pip pipeline + percolator I was forced to add the searchengine pipleine on the config fille too, don't know if it's intended.

Beside that it works very well and gave me a lot of identifications while using a large database.

Thanks!, Noé

[RalfG]

Hi Noé,

We have just released MS²PIP v3.8.0 and MS²Rescore v2.0.0. Both include numerous fixes.

Also if I wanted to use only the Ms2pip pipeline + percolator I was forced to add the searchengine pipleine on the config fille too, don't know if it's intended.

In MS²Rescore v2.0.0 you can fully customize which feature sets to use. I would highly recommend to use all available feature sets (["searchengine", "ms2pip", "rt"]), as all sets are complementary to each other and each can give its own boost to Percolator's sensitivity.

With regard of the modification label issue, are the long-format labels (e.g. Oxidation (M)) a custom setting in MaxQuant, or is the default in a newer version? If the latter, which version of MaxQuant did you use?

Best, Ralf

[RalfG]

Hi @NoeGuilloy,

As of v2.1.1, we now support the long-format modification labels in MaxQuant msms.txt files. Let us know if you would experience other issues.

Best, Ralf