consistency in file output between programs

[saratenhave]

I don't understand why you would generate an out put from Xisearch which cannot be directly input into xiSpec, the headers are inconsistent and you have to reformat the entire file- one column needs to be a score and there are several scores out put from Xisearch etc.
Additionally, why does Xisearch accept .apl files but xiSpec doesn't? therefore you cannot view the output of xisearch at all, these inconsistencies seem like a massive usability over-sight.

[colin-combe]

I made an issue on the xiSPEC parser project about supporting .apl files
(Rappsilber-Laboratory/xi-mzidentml-converter#11)

xiFDR's output is more inline with the xiSPEC input than xiSearch's output

@lutzfischer may be able to update you on whether it is planned to change the CSV output of xiSearch

The xiSPEC website also accepts mzIdentML, which is the actual standard for this type of data.

[lutzfischer]

@saratenhave The output of xiSearch will be modified to be more inline with what xispec requires. But you can use xifdr just as a converter - i.e. read in the xisearch result; set in complete setting minimum peptide length to 1; fdr to 100 and write out the result.
Similarly you could a use a small tool that comes with of xiSearch to convert apl-files to mgf-files.
You can start it by running

java -cp XiSearch[version].jar rappsilber.gui.localapplication.ScanFilter

in the same folder as the XiSearch jar. Replace XiSearch[version].jar with what actual jar-file (i.e. the version you are using).

If you do it before the search you can convert all apl-files into a single mgf (use the second tab to select the apl-files).

Post-search this will only be helpful if you did not search the *.peak.apl files as there the index of spectra can be different in the mgf compared to the apl-file. But otherwise you can convert each file individualy and then rename the entries under peakfilename to the mgf files

@colin-combe is the xiSpec csv-parser the same as in xiview?

[colin-combe]

Thanks Lutz.

I was thinking converting the apl files before might be a better solution.

@colin-combe is the xiSpec csv-parser the same as in xiview?

yep, it's the same

[lutzfischer]

@saratenhave sorry to elaborate a bit more. But my suggestion would be to convert your apl-files before the search to mgf and feed that to xiSearch.

As otherwise you might have troubles with the difference of how mgf and apl files handle spectra with unknown charge states.

We are working on implementing some of the advantages we get from maxquant-apl files in an alternative way. Then msconvert based mgf-files become the de facto standard for xi-search.

[saratenhave]

Thank you for the prompt and comprehensive responses- I will try everything you have suggested and let you know how I go.
Thanks
Sara

[saratenhave]

Hi Lutz,
I don't have the file you suggested should be in the Xiserach folder (java -cp XiSearch[version].jar rappsilber.gui.localapplication.ScanFilter), is it in a previous version? (I have got 1.6.742)
thanks
Sara

[lutzfischer]

@saratenhave
This is part of the XiSearch.jar
so you need to run this in the same folder as xi is.

actually it is
java -cp XiSearch.jar rappsilber.gui.localapplication.ScanFilter

so you need to open a terminal (I.e. start cmd) and change to the right folder and then run the above command.

I attached a bat-file - SpectraFilter.zip -(had to compress it to zip - otherwise it is not accepted here) that should be in the same folder as XiSearch.jar
meaning uncompress the zip and place the
SpectraFilter.bat
in the same folder as the XiSearch.jar
then you should be able to just start the SpectraFilter.bat
and that should show you the gui for this.

Sorry that I was a bit unclear in my reply before.

I would suggest to first convert the apl-files and then search the generated mgf-file.