Hi Mike,

Thank you very much for your quick reply!

I combined the two result files, in this way I got 240 results, but the 
same problem appeared again. Do you have any other suggestions for 
validator software that works well with Kojak?

Thanks a lot,

Gabriella


On Monday, November 12, 2018 at 9:39:01 PM UTC+1, Michael Hoopmann wrote:
>
> Hi Gabriella,
>
> I’m not an expert on Percolator (or machine learning) but I think the 
> problem is simply that the dataset is too small. There are only 128 
> results, which need to be split into training and cross-validation test 
> sets, and thus there isn’t much data to train or test the models. Manually 
> looking at the results show some PSMs to be highly likely to be correct, 
> but Percolator requires a lot more PSMs to perform its analysis correctly 
> (a limitation of this type of validation approach).
>
>  
>
> An option you have would be to combine your intra and inter protein Kojak 
> results into a single dataset to give Percolator. Likely, you have a lot of 
> intra-protein PSMs. Both of these result sets have the same parameters for 
> Percolator and thus they can be combined, even if it is not the most ideal 
> approach. Combining intra and inter protein PSMs can be done with 
> cut-and-paste in Excel or a text editor. Just make sure you have only one 
> header line at the top of the file.
>
>  
>
> Cheers,
>
> Mike
>
>  
>
> *From:* [email protected] <javascript:> [mailto:
> [email protected] <javascript:>] *On Behalf Of *Gabriella 
> Gellen
> *Sent:* Thursday, November 08, 2018 7:15 AM
> *To:* spctools-discuss
> *Subject:* [spctools-discuss] Percolator stopped working using Kojak 
> result files
>
>  
>
> Hi Everyone,
>
>  
>
> I'm quite new to Percolator and now I am practicing with data, that has 
> already been published to have crosslinks using Hekate (
> https://www.ncbi.nlm.nih.gov/pubmed/24010795). I am trying to validate 
> the crosslinked results found by Kojak 1.6.1 using Percolator 3.02. 
> However, the program stopped working, and I got no results. It says 
> "Warning: No target-decoy protein pairs found for the picked protein 
> strategy.
>
>   Check if the correct decoy pattern was specified with the -P flag;
>
>   if the target protein is called "protA" and the decoy protein 
> "decoy_protA", use the option "-P decoy_"."
>
> But the settings for the decoy pattern were correct.
>
> What do you think can be the problem for this?
>
>  
>
> Thank you for your help in advance,
>
>  
>
> Gabriella
>
> -- 
> 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 [email protected] <javascript:>.
> To post to this group, send email to [email protected] 
> <javascript:>.
> Visit this group at https://groups.google.com/group/spctools-discuss.
> For more options, visit https://groups.google.com/d/optout.
>

-- 
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 [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/spctools-discuss.
For more options, visit https://groups.google.com/d/optout.

Reply via email to