Protein level FDR

[fredrik]

The current protein assembly function will not attempt calculating FDR at the protein level. It would be useful to perform FDR estimate using the target-decoy principle. Functionality should be when this is enabled: 1) Only use peptides passing the FDR cutoff (as now) 2) Look for decoy prefix in the proteins matching the peptides. If a peptide matches both target and decoy prefixes, the decoy proteins should not be brought forward. 3) Calculate scores by multiplying the FDR of all peptides for all protein. 4) Estimate protein FDR by counting better decoy proteins at all score estimates. 5) Set protein FDR of target proteins 6) Filter protein list so that only proteins with better FDR are reported.

[fredrik]

(In [3756]) Refs #672. Added protein FDR calculation and filtering. Uppdated form accordingly.

[fredrik]

(In [3757]) Refs #672. Updated plug-in info.

[fredrik]

(In [3761]) Fixes #672. Updated help text

[fredrik]

(In [3774]) Refs #672. Corrected form.

[fredrik]

(In [3910]) Refs #672. Avoiding rare null pointer exception

[fredrik]

(In [3912]) Refs #672. Handling spectrum lookup problems.

[fredrik]

(In [3913]) Refs #672. Handling peptide total intensity of null

[fredrik]

(In [3914]) Refs #672. Handling peptide total intensity of null

[fredrik]

(In [4113]) Refs #672. Writing precursor quantity of generated proteins based on included peptides.