search

vdemichev / DiaNN

DIA-NN - a universal automated software suite for DIA proteomics data analysis.
Other
283 stars 53 forks source link

MS1 Calibration Error #683

Open jsalbin1 opened 1 year ago

jsalbin1 commented 1 year ago

I have been using the DIA_SpecLib_Quant workflow for GPF-DIA data in FragPipe recently, largely without changes other than specifying expected mass shifts. Everything seems to work fine, except when it gets to the DIA-NN quantification step, at which point I always get an error that looks something like this:

[0:27] File https://github.com/Nesvilab/FragPipe/issues/1/8 [0:27] Loading run C:\Users\Owner\Desktop\Pentelute\LCMS_Data\352_Orbi_Start_23_04_09\352\23_04_24\352_23_04_24_DIA_4312_1.mzML [0:59] 45593 library precursors are potentially detectable [0:59] Processing... [0:59] RT window set to 1.0056 [0:59] Peak width: 2.852 [0:59] Scan window radius set to 6 [0:59] Cannot perform MS1 mass calibration, too few confidently identified precursors [0:59] Using MS1 mass accuracy: 20 ppm [1:00] Optimised mass accuracy: 7.02463 ppm Process 'DIA-NN' finished, exit code: -1073741819 Process returned non-zero exit code, stopping

I have attached the full log file from this particular run. It seems to be something related to MS1 calibration, but will persist even if I add "--no-calibration" to the command line inputs (or any of several other command line options that I have tried). It happens both with peptide libraries and with tryptic digests that I have run with the same methods on the same Lumos. Could you please guide me on how to get past this error?

log_2023-04-29_21-47-01.txt

jsalbin1 commented 1 year ago

I've been looking at this more. It seems like it might be some kind of file system error rather than a calibration issue. I've tried running from command line as administrator, and am getting the same error. When I compare the logs in the MSF-DIA preprint and mine, there aren't many differences, but I bolded a few. Could any of these be causing the error, or is there some other known way around this?

Preprint MSF-DIA (HeLa) DIA-NN [Work dir: /storage/yufe/results/msfragger_dia_paper/2018_narrow_wide_hela/msfraggerdia] /storage/yufe/tools/FragPipe-18.0/fragpipe/tools/diann/1.8.1/linux/diann-1.8.1 --lib library.tsv --threads 56 --verbose 1 --out diann-output.tsv --qvalue 0.01 --matrix-spec-q 0.01 --matrices --no-prot-inf --smart-profiling --no-quant-files --peak-center --no-ifs-removal --predictor --dl-no-rt --dl-no-im --strip-unknown-mods --cfg /storage/yufe/results/msfragger_dia_paper/2018_narrow_wide_hela/msfraggerdia/filelist_diann.txt DIA-NN 1.8.1 (Data-Independent Acquisition by Neural Networks)

Mine: DIA-NN [Work dir: C:\Users\Owner\Desktop\FragPipe_PsA_Pilot_Special_Folder_23_05_02\Experimental_Output] C:\fragpipe\tools\diann\1.8.2_beta_8\win\DiaNN.exe --lib C:\Users\Owner\Desktop\FragPipe_PsA_Pilot_Special_Folder_23_05_02\Experimental_Output\library.tsv --threads 11 --verbose 1 --out diann-output\diann-output.tsv --qvalue 0.01 --matrix-qvalue 0.01 --matrices --no-prot-inf --smart-profiling --no-quant-files --peak-center --no-ifs-removal --predictor --dl-no-rt --dl-no-im --strip-unknown-mods --fixed-mod UniMod:39,45.987721,C --cfg C:\Users\Owner\Desktop\FragPipe_PsA_Pilot_Special_Folder_23_05_02\Experimental_Output\filelist_diann.txt DIA-NN 1.8.2 beta 8 (Data-Independent Acquisition by Neural Networks)

jsalbin1 commented 1 year ago

Also, it occurred to me that I should try standalone DIA-NN 1.8.1 with the FragPipe library. This seems to work. I'm wondering whether there is a relevant difference between 1.8.1 and 1.8.2 beta 8 or something about how DIA-NN is implemented in FragPipe that might be causing the issue when run through FragPipe.

vdemichev commented 1 year ago

Not sure what beta 8 is. All beta versions are distributed for collaborators to enable some specific functionality, but they are definitely not thouroughly tested.

This does seem like some kind of bug. If it only happens for 1.8.2 beta 8, then all fine (i.e. it can be buggy). If you also see the bug with 1.8.1, I would be grateful if you could then share library.tsv that FragPipe passes to DIA-NN as well as 352_Orbi_Start_23_04_09\352\23_04_24\352_23_04_24_DIA_4312_1.mzML.

jsalbin1 commented 1 year ago

It seems to work fine if I load the FragPipe library into standalone DIA-NN 1.8.1. I'm happy to share the files if that will be helpful. Or is this something that I should take back to the FragPipe board?