[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [modeller_usage] changing the seed



Amanda Moraes wrote:
> My name is Amanda and I`ve read that changing modeller seed helps to
> achieve a better model.

It will give you a *different* model, not necessarily a "better" one.
(For regular homology modeling runs, setting the random seed is
essentially equivalent to setting ENDING_MODEL to build multiple
different models.) The most obvious way to get a better model (given the
same starting alignment and restraints) is to build multiple different
models and then pick the best one using some assessment method.

> In my top file, where can I put the new comand?

If you're using TOP, you can put it anywhere in the script after the
INCLUDE but before the CALL.

> And how to write it?

For example:

SET RAND_SEED = -4000


However, TOP is very old - you'd be better off using a Python script
with modern versions of Modeller. The equivalent Python script for your
modeling run (complete with a different random seed) would be:

from modeller import *
from modeller.automodel import *

log.verbose()
env = environ(rand_seed=-4000)

a = automodel(env, alnfile='novo_alinhamento.ali',
              knowns=('1MDY', '1NKP', '2QL2', '1HLO'),
              sequence='twist1')
a.starting_model = 1
a.ending_model = 100
a.make()

	Ben Webb, Modeller Caretaker
-- 
             http://www.salilab.org/modeller/
Modeller mail list: http://salilab.org/mailman/listinfo/modeller_usage