From owner-biophysics@net.bio.net Mon Mar 07 22:00:00 1994 Path: biosci!daresbury!trane.uninett.no!nntp.uio.no!usenet From: dave.buss@studorg.uio.no (Chuck An) Newsgroups: bionet.biophysics Subject: Re: Protein Structure Prediction Date: 8 Mar 1994 14:09:48 GMT Organization: NASA, Kennedy Space Center Lines: 27 Message-ID: <2li0vc$4ri@hermod.uio.no> References: NNTP-Posting-Host: biotek16.uio.no X-Newsreader: WinVN version 0.80 In article , varma@lilac.cs.ucdavis.edu (Hemant Varma) says: > > Is any one out there interested in Protein Structure Predictions >from amino acid sequences? I mean the theoretical methods, >molecular dynamics, statistical mechanics.. etc. I am currently >exploring the use of Artficial Intelligence methods in the form >of Artficial Neural Networks and Genetic Algorithms in the prediction >of protein tertiary structure predictions. > >Regards, > >Hemant > I would be very interested in prediction of protein structures based on the sequences. However, I have noticed that none of the currently existing arigorithm has an average accuracy beyond about 66%. It probably indicates unless a better arigorithm is well established theoratical research regarding the prediction of protein structures won't have any break through. No matter how intellegent a system is, its results won't beyond the arigorithm the system is based on. However, regarding the capacity of a work station, the higher the capacity( or speed ), the better the results only when some sort of mimic arigorithm is applied. Just mean to be some personal oppinions. It would be great pleasure to hear different oppinions. Chuck An .