Closed MatteoLacki closed 3 months ago
Just a follow up: the same error met on v3.0.0-b1.
The matter was resolved by using precisely the m/z entries from unimod, i.e. by replacing: "modification_mapping": { "+15.9949":"U:35", "+57.0216":"U:4", "+42":"U:1" }, with: "modification_mapping": { "+15.994915":"U:35", "+57.021464":"U:4", "+42.010565":"U:1" },
I mean, it resolved my problem, but it sure still looks fishy to me.
Altough I got resuls I was getting a lot of errors in the log htat I did not see. The issue is still valid:
I am trying to parse SAGE output. We had a pair coding session with @theGreatHerrLebert which involved moding the results.sage.tsv
to contain direclty proper unimod codes, but this did not work either.
I think we need some more documentation on how to deal with the issue.
Best wishes,
Hi Matteo,
I think (just a hunch, would have to check), that your second attempt with using the exact unimod masses actually just does not rename the modifications, so they remain unresolvable to DeepLC.
In terms of the actual issue, it might have something to do with recent changes in pyteomics.proforma
(https://github.com/levitsky/pyteomics/pull/129), but I would have to verify. Could you perhaps send me (a section of) the PSM file you are using?
Thanks! Ralf
The matter was resolved by using precisely the m/z entries from unimod
That's definitly strange. The keys in the mapping should exactly match what is in the PSM file, so in this case what is reported by Sage.
As I thought, the error is linked to recent changes in Pyteomics. I opened an issue (levitsky/pyteomics/issues/132). In the meantime, you could install an older version of Pyteomics (e.g., 4.6.2)
I could, but maybe rather you could put a limit on its version in your setup.py equivalent?
Point is that then you won't have any more people bothering you in this problem. Do you have a minimal requirements.txt
with a working example?
We indeed try to define dependencies that should work in the pyproject.toml. However, in this case, the problem was fixed in a patch release of Pyteomics (4.6.3), so the pip dependency resolver should always prefer that version over 4.6.2.
Fixed in levitsky/pyteomics#133
Hello,
likely it is something minor, but I simply don't know how to solve it:
I get this error:
with the following config:
version
3.0.0b5
fresh from github.Any ideas?
Best wishes,
Matteo