Julianus Pfeuffer
Julianus Pfeuffer
> nice draft! needs some cleanup though and resolving the discussion regarding use of strings (the standard way in popular python libraries) like scikit-learn. Could this string version only be...
I think we could add an an IM range but it will be quite slow since it will go through a spectrum linearly and check all IM values in the...
I still think that subclasses like TIMSExperiment, FAIMSExperiment, MSExperiment would be worthwhile for efficient handling. But I remember that there were concerns from other sides. They guarantee that things are...
Yes, looks like they are using a sparse offset representation though. Not a dense float array representing every peak.
We need to add it to the pipeline as configs. We can use test_full_lfq, and test_full_tmt. We then should additionally include test_full_lfq as test_full for nfcore compatibility.
How long does it run?
I would say this completely replaces a IsobaricAnalyzer - IDMapper- FileMerger - ProteinInference - ProteinQuantifier workflow as in quantms. A bit less functionality (no unidentified quantities, no normalization). Otherwise the...
> At least for the "unidentified quantities" I know that this feature is sometimes used with unknown PTMs / sequences. So maybe one needs to consider to optionally add this...
You would not. You would re-write it.
For such a project, I think one should first re-implement a simpler FFID from scratch in a way that is easier readable and parallelizable and preferably does not use SWATH...