InstaNovo response to .mzML and .mgf files without MS1 charge states

[vtsiamis88]

Hello InstaNovo Team,

Today, I ran some files that I initially believed were DDA, and I noticed that InstaNovo did not process them as expected. I later learned that the files were actually DIA, which of course do not contain MS1 charge state information.

But there are other cases where DDA files may not include or record MS1 charge state information, and it seems that InstaNovo currently does not provide the user with a clear explanation when this occurs. In my testing, I tried both .mgf and .mzML conversions of Thermo .raw files and observed different responses from the software.

More specifically:

• For .mgf files, InstaNovo issues a warning and filters out all spectra, assuming their charge is >10, which is not the case.

• For .mzML files, the software raises a Python error, treating the precursor_charge value as non-numeric.

I have attached the relevant response files as mgf.txt and mzml.txt, respectively, for reference. Of course other formats might have different response, but I haven't tested.

It would be helpful if the CLI could provide a clearer and more informative feedback when MS1 charge state information is missing, such as in low-resolution DDA or DIA files.

Thank you in advance.
Vasileios

[BioGeek]

Hello Vasileios,

Thanks for the feedback, we'll add better/more informative error messages in a future release.

[tentrillion]

I had problems running instanovo on some mzML data as well. My error message was the same as @vtsiamis88 provided above for mzml.txt.

I was able to get instanovo to run on some different data.

In my case, the data that didn't work was recorded on a Bruker timsTOF and converted to mzML with TIMSconvert. The data that did work was recorded on a Thermo Orbitrap and converted to mzML with ThermoRawParser. I haven't inspected the mzML generated to understand why one worked and one didn't, but I think in addition to (or instead of) better error messages, just some improved documentation might be an even easier fix.

• has instanovo been shown to work on timsTOF data?
• what are the constraints on mzML files that will work? I assume they need to be centroided? (Or no?). What tools (MS Convert, timsconvert, ThermoRawParser) are known to generate appropriate mzML with default settings? (Or at least, is there an XML header that I can grep for in the mzML to see if it will likely work with instanovo?)
• etc. etc.

[tentrillion]

(Note for anyone else who comes across this issue: FileFilter -in my.mzML -out my.filtered.mzML -peak_options:rm_pc_charge 0 from openMS can easily identify and remove scans that are missing precursor charges, and in my case at least this was <2% of the scans for my timsTOF data.)