.d to mgf conversion for de novo about casanovo HOT 6 CLOSED

Can you post a sample of one of your mgf files, or even just the first 1000 lines of such a file?

from casanovo.

The problem seems to be that your MGF file is missing some information, in this case the precursor charges of the MS/MS spectra. Unfortunately I'm not sufficiently familiar to know whether it is expected that this information is not available for timsTOF data, but for successful spectrum identification Casanovo does require it.

The most straightforward solution would be to figure out how to include the precursor charges in the MGF export, although unfortunately I can't directly recommend the best tool to do this.

from casanovo.

My guess is that it is charge is called something else in @jonathan-krieger-bruker's file. A sample of the file's content should help us out.

from casanovo.

Hi all - thanks for helping to look into this. I tested 2 MGF files created different ways but they both look quite different then the test file or preprocessed data provided in the repo. I attach a zip file of these 2 MGFs (Top ~1000 lines or so).
Thanks again!
Jonathan
MGF_samples.zip
.

from casanovo.

The Charge state seems to be there and seems to be written in the same way as the test data. However the intensity values are very different - one looks relative and one looks absolute?

from casanovo.

Relative vs absolute intensity values shouldn't make a difference because intensities are normalized by Casanovo anyway.

In fact, I could run Casanovo successfully on the two example files you provided. I'm guessing that not all spectra are incorrect, and the one Casanovo chokes on is further down in the file you're trying to process. Would it be possible to share the full MGF with which you got the error so I can further debug the issue? Thanks.

from casanovo.