Codes for DeepFLR(paper).

Model construction was performed using python (3.8.3, Anaconda distribution version 5.3.1, https://www.anaconda.com/) with the following packages: FastNLP (0.6.0), pytorch (1.8.1) and transformers (4.12.5).

Data analysis for FLR estimation was performed using python (3.8.3) with the following packages: pandas (1.0.5) and numpy (1.18.5).

For model training, see ReadMe.md in model folder.

Target and decoy phosphopeptide lists of singly phosphorylated peptides are generated by

python Targetdecoy_phosphopeptides_generation_mono.py --inputfile msms.txt --outputfile sequence.csv

where msms.txt is the result from Maxquant.

To deal with doubly phosphorylated peptides, target and decoy phosphopeptides are generated by

python Targetdecoy_phosphopeptides_generation_multi.py --inputfile msms.txt --outputfile sequence.csv

To get DeepFLR FLR estimation, python DeepFLR_result_processing.py --modelresultfile modelresult.csv --sequencefile sequence.csv --outputresult outputresult.csv --outputfileFLRPSM outputfileFLRPSM.csv

Here, modelresult.csv is the result from model/mgfprocess.py, sequence.csv is the output from Targetdecoy_phosphopeptides_generation_{mono,multi}.py.

To get Maxquant result analysis,

python MaxQuant_result_processing.py --inputfile msms.txt --outputresult outputresult.csv