• Sign In
  1. Requests

Two friendly suggestions

support

View Request

view list print
Two friendly suggestions Nathan Manes  2013-05-27 15:18
 
1:
Sometimes I run a peptide-concentration-curve series of replicates, and I get down to the limit of detection. Say I am targeting a peptide using 6 transitions, and that I have a replicate with a peptide abundance near the limit of detection, and say three of the transitions are useful, and the other three are too noisy and not used for quantitation. In the chromatogram of this precursor for this replicate, the noisy three transitions often will obscure the useful three, but I don't want to delete the noisy three transitions from the "Targets" pane (on the left) because they are useful for other replicates that have a greater concentration of the peptide. I suggest adding an option to only display the chromatograms of the transitions that are being used (for the replicate being viewed). Transitions that are not being used for this replicate wouldn't be deleted, just "hidden".

2:
I run synthesized peptides when performing assay development. I suggest allowing users to add an annotation to replicates indicating if they are just peptide standards. If so indicated, then these relative transition intensities would replace the spectral library data when calculating R-squared values.

Related: When importing results, Skyline attempts to identify the peptide peak. Is the R-squared value and the predicted elution time used for peak picking? I think so but it's hard for me to tell. If so, it might be good to make these factors stronger in the peak picking algorithm. I've seen some very strange peaks picked when the correct peak was "obvious", even with the latest version of Skyline (64-bit 1.4.0.4421).

Thanks, and see you at the Skyline meeting at ASMS.
 
 
Brendan MacLean responded:  2013-06-01 12:13
Hi Nathan,
Not really sure how #1 would work in practice. So, I can't say it is likely to get implemented any time soon.

On #2, we are working on implementing the ability to collect targeted chromatogram results and build a "chromatogram library", similar to a spectral library, only composed of chromatograms and peak areas collected using targeted techniques.

We have a prototype working now that allows you to collect Skyline documents into a folder on a Panorama server (http://panoramaweb.org) and then download these collected results as a chromatogram library, which can be added to Skyline in the same way you would add a spectral library. Then the peak area data in the library would server the same purpose as the peak intensities in a spectral library, yes, for choosing the best responding transitions and calculating dot-products (not R-squared) between new measurements and library measurements. If you go further, and use synthetic proteins (as described in this paper - http://www.nature.com/nmeth/journal/v8/n12/abs/nmeth.1770.html) and not just synthetic peptides, then you will also be able to use the peptide peak areas to rank peptides in a protein, which will allow Skyline to pick the best responding peptides for measuring a protein of interest.

And, when you select a peptide in skyline the Library Match view, currently named MS/MS Spectrum view will show you the stored chromatograms.

All this will be supported in the next major release of Skyline due out this summer.

It will also be described in greater detail with screenshots of the prototype code in action at the Skyline User Group Meeting at ASMS and the Panorama poster (MP 382).

The neither the dot-product nor the predicted retention time values are currently used in the Skyline peak scoring algorithm. We are however working on implementing the mProphet semi-supervised learning algorithm to allow us to include new features/scores in our peak picking. This also will be presented at ASMS (TP 409).

Thanks for your suggestions and interest in improving Skyline.

--Brendan