Extra long hours for De-Isotoping of .d folder of TIMs TOF DDA analysis

[JianxiangGegenWang]

Describe the bug It takes 120 to 150 hours per file to finish the deisotoping step. Is this normal? I will also post this in the gmail group. Also another post here might be of similar problem? (https://github.com/cox-labs/CoxLab_Bug_Reporting/issues/21)

MQ ver. 2.4 Raw data: Bruker Tims-tof DDA .d folder (size approx. 6 GB)

Settings:

• Group parameters: Type-TIMS-DDA; Instrument-BRUKER TIM; Lable free quantification tab-LFQ
• Global parameters: Reference Sequence-UP000005640_9606.fasta and UP000005640_9606_additional.fasta;
  Protein Quantification tab, Peptides for Quantification - Unique+razor; Identification - Match between runs

The CPU has 24 cores and I put all of them in the Threads.

Expected behavior Shorter time since if it crashes in the process then it's another 5 days to wait.

Desktop (please complete the following information):

• OS: Windows 10 Pro for Workstations 64-bit
• Version 22H2
• RAM 128GB

Additional context I doubled checked the changes from MQ ver. 2.2 to 2.4. Something changed that might be of interest for this problem:

• Goup Parameters: LFQ tab: LFQ min ratio count changed from 1->2 in Ver. 2.4 Instrument tab: Bruker Tims, Isotope time correlation 0.6->0.3
• Global Parameters: Match between runs: Match time window changed from 0.7->0.4

Thanks in advance for looking into this. Looking forward to hearing from you.

[leonardblaschek]

Just to confirm this issue in a different environment:

Bug: Deisotoping 36 samples would take (by extrapolation) roughly a month of runtime.

Expected behaviour: Expected runtime given sample type, number and hardware should be 1-2 days at most.

Setup: MQ version: 2.4.2.0 and 2.4.9.0 Data: Bruker Tims-tof DDA .d

OS: Fedora Linux 38 (KDE Plasma) x86_64 CPU: Intel i7-10700 (16) @ 4.800GHz Memory: 32 GB

Ran 12 threads to make sure I wouldn't run OOM. Ran through both mono and dotnet to see if that would make a difference.

Update: Using version 1.6.17.0 from bioconda with the same parameters and system, deisotopng took a couple of minutes, so this seems to be a hiccup specific to 2.4.x.