I am not sure that I understand your question.

This is the web page which says for each type of peptide search engine which files you need in order to build a BiblioSpec spectral library:
https://skyline.ms/wiki/home/software/BiblioSpec/page.view?name=BlibBuild

That page says that for Proteome Discover you need the .msf and/or .pdResult files.

If you have peptide search results, the usual place that you start in Skyline is with:
File > Import > Peptide Search

I am not sure whether you are saying that you were not able to get Proteome Discoverer to work, and you would actually like to use Skyline to do your peptide search.

If you want Skyline to perform a peptide search on your raw data, the menu item to do that is also "File > Import > Peptide Search". On the first page of the peptide search wizard you would choose "DDA raw" or "DIA raw" in the "Start From" dropdown if you want Skyline to perform a peptide search.
-- Nick

Hi Nick,
Sorry that I was not clear. What I'm trying to do is to quantify about a dozen of peptides identified in PD across a set of samples. But couldn't figure out how to build the spectral library from the tutorial. My PD search result has >2000 proteins that I don't care about. So importing the entire .msf and .pdResult sounds too much. Then my search has both Sequest HT and Byonic (which is too eager to give me a longer list). On top of that, my raw data was acquired with both target MS2 and DDA, so for some peptides of interest, I have thousands of PSMs. I tried to export selected PSMs from PD to import to Skyline, was told skyline doesn't take PepML from PD. It doesn't help that I've known about Skyline for years but haven't used it at all. Would you help with this?
-Yan

Hi Yan,
If you already know the set of specific peptides you are interested in, you can add them to your Skyline document, then select the "Filter for document peptides" option to build a spectral library using only PSMs with those peptides.
Another option if you just want a smaller set of higher confidence peptides from the entire result set would be to increase the cut-off score when building the spectral library. This excludes results based on score, depending on the source of the results (e.g. cut-off score of 0.95 excludes >0.5 q-value with Percolator results).

Thanks,
Kaipo