I've been using MODELLER 2 ways : first with only one template second with 2 templates Is the Objective Function of the modeles generated comparable ? and if so, how ? The first is around 600, and the second around 3000 for a 151 residus query sequence.
Thanks in advance for your help,
___________________________________________________________________________ Découvrez une nouvelle façon d'obtenir des réponses à toutes vos questions ! Profitez des connaissances, des opinions et des expériences des internautes sur Yahoo! Questions/Réponses http://fr.answers.yahoo.com
doppelt olivia wrote: > I've been using MODELLER 2 ways : > first with only one template > second with 2 templates > Is the Objective Function of the modeles generated comparable ? and if > so, how ? > The first is around 600, and the second around 3000 for a 151 residus > query sequence.
The objective function is the sum of all restraints, and so is only comparable if you have the same number and type of restraints, which is not the case here, so no, they are not comparable. You would be better advised to evaluate the models with an assessment method such as GA341 or DOPE (which will be comparable between models since your target sequence is the same). I would also advise you to build multiple models to ensure adequate sampling, not just a single one for each case.
Ben Webb, Modeller Caretaker
I am modeling a structure composed of two modules: M1 (one domain) linked to M2 (two domains). M1 - 100 residues M2 - 200 residues
First, separately I generated models for M1 and M2 and I joined them. Secondly I generated models for M1 and M2 in this way that they have been already joined - I created one chain with already joined M1 and M2. I separated M1 and M2 and I compared DOPE scores.
Models modeled "alone" had better score: M1 = DOPE = -9360 M2 = DOPE = -20497
than models after "separation"
M1 = DOPE = -9129 M2 = DOPE = -20211
1_Does a difference in DOPE score is significant ? 2_Which models chose ? 3_Which modeling procedure is in better agreement with protein folding problem? Domains fold indepedently but from the other hand a final structure is composed of two joined elements = energy of "final protein" should be influenced by energy of M1 and M2 - so maybe despite worse DOPE scores a chain modeled with already joined domains is "much" biologic.