Which is the fairest secondary structure prediction algorithm of them all?
The importance of secondary structure prediction to bioinformatics, modeling and structure prediction cannot be overstated. In a somewhat recent paper by Palopoli et al., their group found that combining the results of mutliple algorithms (with their JSSPrediction methodology) gave major improvements in prediction accuracy (see table at right). This ‘combine and conquer’ strategy is a tried and true strategy for multiple prediction problems in biological systems.
With this poll, we’d like to see who you go to for your secondary structure predictions and hear your thoughts. Are there algorithms not in the poll that you like? Do you combine algorithms? Do you like the servers or prefer in-house command-line installs? What are the challenges? Is an 88% prediction accuracy the ceiling or will we one day get to 95%?
For Rosetta, the standard is a mix of PsiPred, Jufo, SAM, and Prof.
Loading ...Palopoli, L., Rombo, S., Terracina, G., Tradigo, G., & Veltri, P. (2009). Improving protein secondary structure predictions by prediction fusion Information Fusion, 10 (3), 217-232 DOI: 10.1016/j.inffus.2008.11.004
Random Posts
Enjoyed this Post ?
Subscribe in a reader.
Follow us on twitter.
3 Comments »
RSS feed for comments on this post. TrackBack URL
I see you’ve designed a new algorithm: the Principal User Component Analysis Vector Analysis technique. After all, Support Vector Machines are the second-best solution to any problems.
.
Why is SABLE not on the list?
@A.G.Evdokimov – it is now. Thanks for the suggestion and please feel free to add your vote for SABLE.