When using the same raw files and protein sequence data, the detection results differ significantly across different versions.

[mondayz11]

Hello, we used three different software versions—2.1, 2.4, and 2.7—for the search and observed a progressive decrease in the number of proteins detected; while the target protein was successfully identified using version 2.1, it was not detected with versions 2.4 and 2.7.

proteinGroups-2.1.txt
proteinGroups-2.4.txt
proteinGroups-2.7.txt

[mondayz11]

These are the three configuration files.

mqpar2.1.xml

mqpar2.4.xml

mqpar2.7.xml

[WalterViegener]

what instrument did you use?

[mondayz11]

Orbitrap Astral

[WalterViegener]

Dear @mondayz11,

thank you for the report and the parameter files. The three runs weren't processed with the same instrument settings. In 2.7 the instrument type was Astral, but in 2.1 and 2.4 it was Orbitrap. The instrument type controls feature detection and the MS1 settings, so the results from the different versions can't be compared directly.

For now, please set Group-specific parameters → Instrument → Instrument type = Astral in every version you compare, and keep the rest of the parameters identical. The identifier parse rule should also be the same everywhere (2.7 used >(.) instead of >([^\s])). Astral support was added in 2.4.3, so 2.1 can't process this file correctly.

About the 716 protein groups in 2.1: every row in that file has the same Score (323.31). That is a numeric saturation, so the protein FDR had nothing to rank by and let all groups through. The 716 therefore isn't a 1 % FDR result. We have fixed this for the next release. Also, your target protein E13 is in the 2.4 output (7 peptides, 56.9 % coverage).

If the numbers still differ after you re-run with Astral, please send us summary.txt, msms.txt and evidence.txt from each run. Please upload it here.

Best,
Walter