unable to get similar results in TPP 7.1 and earlier versions

[reeju]

Hello everyone,

I am running a raw file in TPP 7.1, which was also analyzed previously in an earlier version via TPP 6.3. I have kept the comet parameters and peptide prophet settings similar for both the cases, still there is difference in proteins identified from both TPP versions. In tpp6.3 the total proteins identified were 1546, after applying filters on the same file with probability 0.88(having 0.05 error rate) the protein count was still 494. But in tpp 7.1 the total protein count displayed was only 99 without any filtering on the same file. Why is there such a drastic difference. Comparison of both files have been attached. If anyone can help us troubleshoot please revert back.

Thank You

Regards

--

Reeju Rani

Research Scholar

Proteomics and Cell Biology lab

ICAR-National Dairy Research Institute

Karnal, Haryana India

[David Shteynberg]

Hello Reeju,

Thank you for trying the TPP for you analysis and for your questions and slides.  Here is what I think might be happening.  In 6.3 the scores from comet and the PeptideProphet analysis based on the models were different from 7.1.  

Here is the comet f-value model in 6.3:

Here is the f-value model from 7.1:

 

The goal of PeptideProphet is to identify a bimodal distribution and fit two distributions to the observed black line, with the corrects modeled in green and the incorrect in red.    In the first case (6.3) the observed distributions is not really bimodal to the eye and the PeptideProphet model is fitting both distributions to the one peak because the observed distribution is not bimodal enough, and please correct me if I am wrong, you are not providing known true negatives (decoys) to help train the model and keep the red distribution “pinned" by them.  In the second case (7.1), the bimodal distribution is fitting much better in my opinion as the red curve is mostly in the correct place and the green model is fitting the shoulder where the majority of the correct PSMs in your data will be.   I know you feel like you “lost” some IDs going fro m 6.3 analysis to 7.1, but I would say you likely improved the confidence in the results by the reanalysis.  You can verify this by incorporating unknown true negatives (entrapement decoys) in you database to use as another FDR estimate to compare to the model’s results from 6.3 and 7.1.

Cheers!

-David

--
You received this message because you are subscribed to the Google Groups "spctools-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to spctools-discu...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/spctools-discuss/CAJKhLXzf3UGPXKrergg-s5VuOtciS2O2Vq18tgu2WncQcB5L4g%40mail.gmail.com.
<comet.params><comet.params><TPP.pptx>

[reeju]

Thankyou for explaining what is happening in this case.

1. I further want to confirm which model I should prefer as my objective is to identify the maximum number of proteins present in the sample.

2. Also how to incorporate decoys in the database.

Regards

Reeju

To view this discussion on the web visit https://groups.google.com/d/msgid/spctools-discuss/FCA9687F-9C6B-45E5-A917-4B400C05F189%40systemsbiology.org.

[David Shteynberg]

Hello Reeju,

The preferred model would be the one to maximize the correct identifications (sensitivity) and minimize the incorrect identifications (error) at the same time.

To generate decoy databases you can use the TPP Tools:Decoy Database option.

Cheers!

-David

On Sep 4, 2024, at 9:08 PM, reeju <ree...@gmail.com> wrote:

Thankyou for explaining what is happening in this case.

1. I further want to confirm which model I should prefer as my objective is to identify the maximum number of proteins present in the sample.

2. Also how to incorporate decoys in the database.

Regards

Reeju

On Mon, Sep 2, 2024 at 8:52 PM David Shteynberg <dshte...@systemsbiology.org> wrote:

Hello Reeju,

Thank you for trying the TPP for you analysis and for your questions and slides.  Here is what I think might be happening.  In 6.3 the scores from comet and the PeptideProphet analysis based on the models were different from 7.1.  

Here is the comet f-value model in 6.3:

<PastedGraphic-1.png>

Here is the f-value model from 7.1:

To view this discussion on the web visit https://groups.google.com/d/msgid/spctools-discuss/CAJKhLXyTqZaXEJW0bFbJ%2BcEx%2B8aEpbAjE7MV4Ttoa3x21hfZyw%40mail.gmail.com.

[reeju]

Hello,

I am analyzing another file for label free quantitation using stpeter. During command it was showing that 82 proteins have sin value, yet the file i am receiving after analysis is empty. 

To view this discussion on the web visit https://groups.google.com/d/msgid/spctools-discuss/20A2DD96-85F1-41C7-B115-4D9C888C64E0%40systemsbiology.org.

[David Shteynberg]

Hello Reeju,

I think what you might be encountering is a known bug in the current version which has the following solution: https://groups.google.com/g/spctools-discuss/c/x004zkrVKjI

Please let me know if that does not work for any reason.  The bug will be corrected in the next release of the software.

Cheers!

David

To view this discussion on the web visit https://groups.google.com/d/msgid/spctools-discuss/CAJKhLXzPV1BiWYqZEmyna--Fb-DiBNO48f5BBDBXw55wO8wV2w%40mail.gmail.com.

[reeju]

Hello,

I have attempted to resolve the issue as instructed above but the problem persists.

To view this discussion on the web visit https://groups.google.com/d/msgid/spctools-discuss/CAGJJY%3D_9yk5Lw4ESBTEnvhvn4-d41ro4Z7UfK3Ac-FJ42DZsuw%40mail.gmail.com.

[David Shteynberg]

You will need to re-export the results again after patching your TPP.   Thanks.

To view this discussion on the web visit https://groups.google.com/d/msgid/spctools-discuss/CAJKhLXxqQp9Nbue45SkMasJM3NJVYDjgz3HLAVcrrAbAHba4eA%40mail.gmail.com.

[reeju]

Hello,

Thanks for the help. The problem has been resolved.

To view this discussion on the web visit https://groups.google.com/d/msgid/spctools-discuss/CAGJJY%3D8S2E53gcZtMCFR9x%3D0eHUOVpqmEteRbdRPdfH29E0%2BdA%40mail.gmail.com.