protein validation

35 views
Skip to first unread message

felipe velasquez

unread,
Apr 30, 2023, 10:13:34 PM4/30/23
to spctools...@googlegroups.com
Hello

I have a question about protein validation. I am trying to use TPP to identify post-translational modifications of certain proteins. I performed a Comet search using the sequence of my protein of interest. However, while exploring another database, I performed a search using a protein that is not in the experimental design, as it comes from a virus. Surprisingly, this protein appeared in my analysis. Is this normal?

thanks,

Felipe Ignacio Velásquez Salinas
Bioquímico
Doctor en Microbiología

Luis Mendoza

unread,
May 2, 2023, 10:53:22 PM5/2/23
to spctools...@googlegroups.com
Hello Felipe,

There may be several reasons why you see the viral protein in your analysis, e.g.
  • Did you only run Comet, or did you validate using the TPP/Prophets?  If the latter, what was the probability of the protein(s) in question?
  • What were the contents of the database that you used to search?  Did they contain only "the sequence of my protein of interest"/virus protein, or did you use a full proteome database?  Did you use decoys?  Contaminants?
  • Does the virus protein contain peptides that may be shared with proteins in the sample itself?   Do you observe unique/proteotypic peptides from that protein?
Cheers,
--Luis



--
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/CAA35R_VO9uJbjkdKoJKrZVFtRnuV04Hp%2B8FMzE1BqaiTYtmx8Q%40mail.gmail.com.

felipe velasquez

unread,
May 4, 2023, 12:41:31 AM5/4/23
to spctools...@googlegroups.com
Hello Luis

Thanks for your request. Yes, I performed TPP/Prophets, and their probability is high as 0.99999 or 1, in many peptides. The database I used only contains my protein of interest in FASTA. I do not know how to use decoys.

Kind regards,

Felipe

Jimmy Eng

unread,
May 4, 2023, 3:30:58 PM5/4/23
to spctools...@googlegroups.com
Felipe,

You should not search sequence databases composed of just a single protein or even tens of proteins and hope to process them through downstream validation tools like the Prophets.  When you do this, you break lots of assumptions that the tools expect which makes the results, like probability of 1 for many peptides, meaningless.  Ideally you search the full organism protein sequence database.  But if you feel you understand the analysis and tools well enough that you need to search a smaller database, minimally use 500 or 1000 protein sequences and include your protein of interest.

Jimmy

felipe velasquez

unread,
May 5, 2023, 11:42:40 AM5/5/23
to spctools...@googlegroups.com
Dear Jimmy

Thanks for your reply. I will use the correct database. 

Felipe

Reply all
Reply to author
Forward
0 new messages