Reading Fasta is slow on importing MetaMorpheus results

support
Reading Fasta is slow on importing MetaMorpheus results rmillikin  2018-06-14 13:33
 

Hello from the MetaMorpheus dev team!

I'm trying to get our .mzID search results to import correctly into Skyline. The .mzID imports successfully with Skyline Daily (but not Skyline 4.1 release, which is fine). However, when I add my .fasta database, the process gets stuck.

To reproduce:
Download .mzID, .fasta, and .mzML from https://uwmadison.box.com/s/4qk0rhjbvcz1jmz6l2ax7vqx0wxxfvpf
In Skyline Daily, Import DDA Peptide Search
Add .mzID, click Next
Spectra get read
Click thru menus to extract chromatograms, select mods, etc
On "Import Fasta" Browse and select fasta, 2 missed cleavages, click finish
Progress bar appears and proteins are added, but extremely slowly. No real progress is made even after >1 hour

Let me know if this is an issue with our .mzID format, etc. We'd be happy to change code to get this to work.
Thanks,
Rob

 
 
Brian Pratt responded:  2018-06-14 15:01

Hi Rob,

It will probably be helpful to have screen shots of any settings you selected along the way. When I use your files, accepting all the defaults, everything proceeds more or less instantaneously.

Thanks,

Brian Pratt

 
Nick Shulman responded:  2018-06-14 15:10

I think the problem is that if you tell Skyline to accept all of the modifications (there are about 60 of them) that were discovered in the peptide search results, then Skyline has too many permutations that it needs to work through when Skyline is trying to figure out what modifications to put on the peptides from the FASTA file.

One thing that you can do instead is cancel out of the "Import Peptide Search" wizard once you get to the part where it's asking for the FASTA file.
Then, in Skyline, use the menu item:
View > Spectral Libraries
and then use the "Add All" button to add the peptides from the spectral library into your document.

I am not sure whether we can speed up all of the work that Skyline is doing during "Import Fasta". It seems like something that ought to be straightforward to fix ("instead of generating all possible peptides and then checking whether they are in the library, instead iterate over the peptides that are in the library") but there might be a reason that it is implemented the way that it is.

 
rmillikin responded:  2018-06-15 09:22

Thanks to you both, indeed when I don't select any mods, the import is very quick. We're writing PTM discovery software though, so we're going to often have a fairly large variety of PTMs observed in a sample.

The cancel and then view->spectral library tricked worked, thanks Nick. Great talking to you at ASMS at my poster, too.

Our search output is largely in .tsv format and we're excited to use Skyline for visualizing our search results (from mzID). It's a great help to us and we're of course fans of open-source software. Thanks for contributing so much to the MS community.