How to annotate unclassified proteins?

Question

[Firingam]

Many of the proteins of P.falciparum's are not annotated, so it is difficult to see if a particular protein is a kinase rather than another protein type. I used Panther Classes to get some data, but many of the Protein ID I gave to it are returned unclassified and I feel my analysis is a bit poor of significance. Therefore, I was wondering if there is any tool to predict what kind of protein I am dealing with (hopefully not to be paid). I thought about a strategy, like finding the orthologues by blast in another well annotated organism , taking these names and then infer. However prior to do this I would like to know if there is something more reliable. Thank you in advance

Answer 1

[Mensur Dlakic]

Many of the proteins of P. falciparum are not annotated because they can't be annotated at present, not for the lack of trying. I don't mean to presume too much, but for a model organism such as this it is unlikely that most people will do a better job with annotation than what is already available. I am saying this to alert you that you may spend a good deal of time trying to annotate these proteins and have little or nothing to show for it.

Also, it depends on what annotating a protein means to you. Usually it means describing its molecular function (e.g., histone lysine methyltransferase), but sometimes even a general biological function will suffice (e.g., involved in cell division). Either way, your proposed strategy will take a long time and likely not work. More recent annotation tools work with databases of protein families such as Pfam. See here and here how to install it locally so you can test your proteins of interest. You may want to try a conserved domains search tool or install it for local use as described here.

Finally, you may want to try the HHpred server, but it is not easy to install it locally. I suggest you try the web interface first and decide if it is worth the effort.

[Joe]

Second everything Mensur has said - those proteins are not annotated for a good reason. HHPred is as good a tool as I have found for dealing with unknown proteins, but it is indeed (or at least was) hard to install.

I have some tips here if you do plan to install it, but they are somewhat out of date now just FYI. A: HHpred batch submission