You are here

conway2016 correction on beta-nov15?

6 posts / 0 new
Last post
conway2016 correction on beta-nov15?

conway2016 correction was shown to be a useful improvement when applied to talaris2014. I just wonder if it makes sense to apply it to the new generation of scoring functions, say, beta-nov15 in Rosetta3.8 ?

Post Situation: 
Wed, 2017-03-15 05:26

I'm pretty sure it's not safe to apply Talaris14 corrections to beta15.  I expect Conway's work is considered in beta15 (don't let the dates fool you - beta15 will have a 17 name when we make it the official scorefunction).  I've asked for clarification.

Wed, 2017-03-15 06:39

Thanks for the quick answer! 

I was more concerned if it is a good idea to apply conway2016 correction to beta-nov15, rather than Talaris14 correction to beta (just curious: how could you possibly do that, apart from writing you own correction file containing Talaris2014 weights? )

Conway 2016 (ProteinSci) does not cite Park2016 (JCTC), neither does Park2016 cite Conway2016

Wed, 2017-03-15 08:49

By "Talaris2014 corrections", I meant "the Conway corrections to Talaris2014".  I'm assuming they are corrections to 2014 - I don't remember off the top of my head.  I'm no longer academic so I can't readily grab the paper to check, and it's not written into the code in such a way that I can grep for it.

Both papers are out of the DiMaio lab, so I think it's safe to assume that beta15 contains the ideas from both sources; I can't comment on why they don't cross-cite.

Wed, 2017-03-15 10:23

Frank (DiMaio) lets me know that the Conway corrections will be "merged" with beta15 for beta16 (which I think is probably available, if unpublicized, in 3.8).  So, don't try to combine them yourself, but do give beta16 a spin.

Wed, 2017-03-15 20:34

Thank you very much for the clarification! 

Indeed, in the help page for beta16, conway16 correction is introduced:


Thu, 2017-03-16 01:21