0
$\begingroup$

I'm trying to use Jpred to predict secondary structure for a protein sequence. When I run J-pred, I get a bunch of hits from PDB. I've also noticed these 'hits' are the same name as the templates i use when using Swiss mod. How do i interpret these hits, do I just ignore them and force jpred to predict structure anyway?

$\endgroup$
0
$\begingroup$

That depends on what your are trying to do. Apparently your query sequence is similar to that of proteins with a known 3D-structure. As it says on Jpred's result page (and help), in this case it might be worth looking at these homologues with experimentally determined structures for information on secondary structure. Most likely it will be more accurate than Jpred's prediction.

How much of the secondary structure you can use from the 3D-structure also depends on which parts align between your sequence and the ones in the PDB. You can show the alignments on Jpred's result page by clicking the button below 'Alignment of PDB hits to your sequence'. However, if you are going with annotations from the 3D-structure, it would probably be better to use SWISS-MODEL directly.

If you don't want to see those hits and just force a Jpred prediction, you can tick 'Skip searching PDB before prediction' in the advanced options before submitting your prediction. The hits are shown merely as a convenience, ignoring them doesn't change anything about Jpred's prediction.

$\endgroup$
  • $\begingroup$ Thank you! So chou &Fasman, Gor, etc are just alternative alogrithms to Jpred? Do you think it would be worth comparing the results of multiple alogrithms or would one be sufficient? $\endgroup$ – Ksims May 3 '15 at 14:35
  • $\begingroup$ @Ksims Yes, those are all secondary structure prediction methods, albeit of varying age: Chou-Fasman is one of the early methods from the 70s, GOR should also fall around that time, while Jpred still seems to be under development (at least until 2000). Wikipedia has a small list of secondary structure prediction methods which you can try out. I do think it is worthwhile to look at a few more results, although the consensus isn't necessarily always right. $\endgroup$ – mivilar May 3 '15 at 15:20
  • $\begingroup$ awesome, I appreciate your help! $\endgroup$ – Ksims May 3 '15 at 18:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.