What Are The Best Protein Beta-Barrel Prediction Tools Currently Available?

Question

[K]

Hi to all.

I am trying to find out if a protein is capable of forming a beta-barrel or not. I have tried several programs, but they seem to give non-additive results (i.e. each one seems to "fight its own war"), or simply not to function anymore. From your experience, which are the most reliable protein beta-barrel prediction tools? (an online server would suit best; I am far from Linux-proficient)

Thanks in advance.

Answer 1

[João Rodrigues]

Quite honestly, a secondary structure predictor should be enough to tell you, depending on the beta-strand content. What do you mean by 'fight their own war'? If they agree that there is a beta-barrel, regardless of number of strands, etc, that's enough of a prediction. They will not predict a structure reliably, at least in my experience.

Here's what we have used in the past in our lab:

• BOCTOPUS
• PRED-TMBB

[K]

In that case, if I do find for instance eight beta-strands, can I confidently expect the protein to form beta barrels? What about the possibility of beta-sheet formation? Regarding the "strange" idiom I wrote, actually it is a cheap translation from a spanish idiom, in this case meaning that the results I obtained by different tools did not seem to be coherent among each other.

Regarding the two tools you suggest: -I tried the first one already, but according to the author, it actually is dependent on having an already predicted beta barrel. What this tool does is assuming that you have a beta-barrel, and predicts if the beta barrel can be transmembrane or not (if I did not get it wrong). -I tried the other tool also. It seems to offer three methods, the two first of which do not seem to give any positive result, and the third one giving results which are incompatible with general secondary structure predictions (JPRED3).

From what you tell me, apparently your experience with beta-barrel prediction has also not been satisfactory, as is my case...

Well, thanks for your reply anyway, João. Regards from Barcelona.

[João Rodrigues]

Hi K,

I understood your idiom (we have the same one across the border!) but I meant if the methods were in fact disagreeing, which apparently they are..

I used the method on something that I knew it was a beta-barrel already.. just wanted to know its topology. Indeed, to predict a beta-barrel from scratch is probably tough.. I would say it depends on the secondary structure and length of your protein. If you see that the majority of your protein adopts beta-strands, then you ought have some sort of barrel-like fold.

Since these prediction methods are not very reliable/coherent, I'd try a different approach. Do you have any functional data for your sequence? What do other proteins in the family look like? Are there homologous structures, even if remotely? Are these barrels?

It's a tricky business..

[K]

João, please accept my late thanks.

I believe I have found a (very) distantly related family of proteins, which actually seems to form beta-barrels. I think I might be able to make a case on all taken together, and at least to give grounds for making strongly plausible that a beta-barrel is really present.

Despite the uncertainty experienced for a very long while, I am enjoying this protein structure study work... too bad I started it too late into my Ph.D.! Saludos desde Barcelona.

[João Rodrigues]

Good it worked out! It's never too late :)

Cumprimentos de Portugal