You are here

Protein Structure Prediction using PyRosetta, suggestions wanted :)

2 posts / 0 new
Last post
Protein Structure Prediction using PyRosetta, suggestions wanted :)
#1

Hello!

Im trying to write a de novo folding script with pyrosetta (PyRosetta
v2.011 for windows).

I'm trying to follow the approach described in this paper:

Alena Shmygelska and Michael Levitt; Generalized ensemble methods for
de novo structure prediction, PNAS February 3, 2009 vol. 106 no. 5
1415-1420

that in turn refers to

Rohl, et all; Protein Structure Prediction using Rosetta, methods in
enzymology vol 383, 2004.

Right now I'm using only low resolution centroid score functions
because I'm not interested in the all-atom refinement stage in my project
(that is something similar to the first paper I cited, i.e. I am
applying replica exchange method to the abinitio protocol)

here I list the main steps of my folding script:

starting from the extended chain configuration

1-I apply 2000 random 9-residue fragment insertions, accepting them
according to the metropolis criterion and using score0

2- 2000 random 9-residue fragment insertions accepting them according
to the metropolis criterion and using score1

3- 20.000 random 9-residue fragment insertions, accepting them
according to the metropolis criterion. Every 2000 insertions i switch
between score2 and score 5

4- 2000 random 9-residue fragment insertions, accepting them according
to the metropolis criteria and using score3

5- 8000 random 3-residue fragment insertions, accepting them according
to the metropolis criteria and using score3

I am not interested in implementing the latest abinitio protocol. I
would rather try to copy exactly the classic abinitio protocol
performed in the first paper.

Is the protocol similar to the default classic abinitio?
Is there something I forgot or that you think I should add?
I will start testing it soon and it would be great if anyone has any
suggestions

Post Situation: 
Tue, 2012-05-15 09:40
oppopomoz

oppopomoz, this sounds about right. Though, there may have been changes in the score function between then and now. Why not use the current executable?

Wed, 2012-05-16 07:15
jadolfbr