In article <1995Jul12.182420.12069 at alw.nih.gov> johnk at spasm.niddk.nih.gov (John Kuszewski) writes:
>> In article <6g7n5o33bw.fsf at hodgkin.mbi.ucla.edu>, arne at hodgkin.mbi.ucla.edu (Arne Elofsson) (Arne Elofsson) writes:
>>> |> On the other side the packing of secondary structure elements is not that
> |> impressive. Looks pretty good for the fragments in fig 3 (1fb) and for
>> I think that part of this is explained by his having "solved" these
> structures in short pieces (because the program is computationally
> expensive).
>
or because it did not work for longer frags ?
> But of course, you're right--the packing of secondary structure elements
> isn't all that amazing. If you looked at the RMSDs of these models
> to the crystal structures, they're pretty high--5 A is about the best
> (if I'm remembering correctly), and some are 10 A away.
>
5 A is very close to random for a compact 50 aa fragment with correct
sec.str. 10 A is certainly random.
> To start another thread, are models of that resolution useful for
> anything?
>
I think they could be usefull for searching a library of folds.
(Actually with 90% secondary prediction I'd bet you I can find
the structure in a fold library given that it is not a new fold).
This is basically what Hubbard/Eddy and Benner did in Assilomar
but with much lower sec.str prediction accuracy.
> |> It is interesting that such a simple method seems to work that well.
>> Precisely. I just saw Andrej Sali give a talk on MODELLER, and its
> output is amazingly good. However, he's using a very large empirical
> database. LINUS does extremely well for having so little starting
> information.
>
Tell ous more. What is new in MODELLER. I did not know it did sec.str
predictions.
Is Andrej at MSI nowadays ? (Making Modeller a commercial product ?)
> |> Unfortunately this is one of many studies that would be extremely
> |> much better if they had done some more rigourous testing, from
> |> what I read in the John Hopkins journal, the run takes over night
> |> or so, which makes me wonder why they do not run this on 50 different
>> George Rose told us at his (now infamous?) NIH talk that it's more
> like 1-2 weeks per "structure."
>Even so, I'm sure they can borrow 10 workstations by someone
(mine for sure) and run 20 structures in a month.
> |> But atleast it gives me a few new ideas and I guess that makes
> |> it a good paper. It is also described (more or less) good enough so
>> Yes.
>> One last question: Are there any other algorithms that predict
> secondary structure as well as LINUS?
>
The published record is (to my knowledge) is 72.2 %
by Salamov & Solovyev
arne
--------------------------------------------------------
From: Arne Elofsson
Email: arne at hodgkin.mbi.ucla.edu
WWW: http://www.doe-mbi.ucla.edu/arne/main.html
