Wrong quantification for extreme ratio in SILAC H/L mixtures #2593
Description
I'm not sure which module is causing this (philosopher or ionquant), but I can see FragPipe providing totally incorrect results from the analysis of a dilution series of a SILAC H:L mix.
I have measured the same dilutions using different acquisition methods, DDA (in black) or DIA (colors), and processed DDA data with FragPipe and DIA data with other software.
Normalization was disabled, and the data evaluation was performed using the combined_modified_peptide_label_quant and combined_protein_label_quant .tsv files.
As you can see, while at the peptide level, FragPipe follows the trend, at the protein level for extreme ratios, the numbers appear to be incorrect
Peptide level:
Protein level:
This also propagates to the accuracy of the quantification, because if I compare the measured ratio against the expected one from the mixing, this is again largely incorrect for FragPipe at extreme H/L mix
Peptide level:
Protein level:
The DDA data themself are fine. When using another software, the numbers follow what I get for DIA, and protein ratios are now as expected
Have you ever seen this?