Dear MSGFPlus developers,
Describe the question or problem
I noticed a change of e-values between runs, when the order of sequences in the database changes. Is this normal?
Details
I am using the following command to analyze an mzML file with MSGFPlus 2024.03.26 (openjdk 21.0.10 2026-01-20):
java -jar MSGFPlus.jar -d db.fa -s in.mzML -o out.mzid -thread 2 -mod mods.txt -numMods 3 -tda 0 -t 10.0ppm -ti -1,2 -m 0 -inst 3 -e 1 -protocol 0 -minLength 8 -maxLength 40 -ignoreMetCleavage 1 -ntt 2 -minCharge 2 -maxCharge 6 -n 1 -addFeatures 1
I performed multiple runs of MSGFPlus and found that keeping the input file and database the same produces identical e-values. However, when I shuffle the database between runs (keeping headers and sequences the same), I can see a small change in the e-values at the scale of +/- 1 E-21.
Have a nice day,
Felix
Dear MSGFPlus developers,
Describe the question or problem
I noticed a change of e-values between runs, when the order of sequences in the database changes. Is this normal?
Details
I am using the following command to analyze an mzML file with MSGFPlus 2024.03.26 (openjdk 21.0.10 2026-01-20):
java -jar MSGFPlus.jar -d db.fa -s in.mzML -o out.mzid -thread 2 -mod mods.txt -numMods 3 -tda 0 -t 10.0ppm -ti -1,2 -m 0 -inst 3 -e 1 -protocol 0 -minLength 8 -maxLength 40 -ignoreMetCleavage 1 -ntt 2 -minCharge 2 -maxCharge 6 -n 1 -addFeatures 1I performed multiple runs of MSGFPlus and found that keeping the input file and database the same produces identical e-values. However, when I shuffle the database between runs (keeping headers and sequences the same), I can see a small change in the e-values at the scale of +/- 1 E-21.
Have a nice day,
Felix