• Sign In
  1. Requests

Suggestion - Multiple libraries use

support

View Request

view list print
Suggestion - Multiple libraries use vdelcourt  2017-10-12 10:27
 
Hi,

As it's my first post on Skyline support, I'd like to thank the Skyline team for providing such amazing program and detailed tutorials to the community
This a suggestion rather than a bug, so I'll go through each steps.

We are beginning a new targeted proteomics project (using PRM) and we decided to generate a new spectral library (AP-MS experiments on target proteins to generate many high-quality PSMs).

After samples were submitted to the MS, RAW files were identified using MaxQuant and msms.txt files were imported into Skyline. In order to build the m/z selection list for targeted experiments on target proteins, I went through every peptides of targeted proteins and took a look at each MS/MS spectra stored in Skyline. I've been quite surprised to see some low quality MS/MS spectra by clicking on different peptides (left bar of skyline). After looking them up in the MaxQuant viewer, I noticed that these particular peptides had better MS/MS spectra than the one Skyline showed me.

After looking around, I noticed that old libraries (from previous projects but generated on the same instrument) were still used (check box checked). After unchecking those boxes (and leaving only the one we just generated), Skyline now shows me peptides with high-quality MS/MS spectra.

I guess that if I didn't uncheck those "old" spectral libraries, confidence on spectral matching (dot products) would have been really low and thus might have been wrongly discarded.

My suggestion would be that Skyline could show best MS/MS spectra (highest score) between multiple loaded spectral libraries. (I think that would be fixed by generating new spectral library with every msms.txt files but this would be also time consuming when importing large spectral libraries)

Best regards,
Vivian
 
 
Brendan MacLean responded:  2017-10-14 17:46
Hi Vivian,
That is an interesting idea, and I will keep it in mind. However, there are also cases where Skyline needs to choose a matched spectrum before you import any data. I think it would be somewhat confusing if Skyline were to change its spectrum choice based on your imported data, and of course the poor matching problem you highlight would still be a problem in these cases.

In general, these libraries are supposed to be best-matching, high-quality spectra for each target for the instrument you are using. In the long run, it may actually be better for you to get the kind of poor results you got that lead you to turn off the other libraries, as that is really the correct solution than for Skyline to try to compensate in this case, but still produce suboptimal results in other cases.

But, thanks again for posting your feedback. It is always useful to know what causes confusion, as that may also lead to other new and creative solutions.

--Brendan