Hi,
I'm using modeller lately, and I haven't had any problems while predicting the tertiary structure of a target using in the multiple alignment sequences with pdb files.
Now that I'm trying to use, only CA files I receive this message and obviously Modeller stops the execution...
> Array too small. Increase MPRMCNS > current maximum, current need: 1393200 1393224
Do you know if this is a parameter I can increase? or what can be causing the error.
Let me explain a little, I'm obtaining these CA files using other methods to find the 3d structure of a given protein X. Then I get obtain all these files X1, X2, X3,... and make a multiple alignment using the same sequence of protein X as the target sequence and the sequences of each X1, X2, X3, that can differ from X in a few residues, as know structures. Then I'm trying to feed all these 3d models of the same protein and the alignment to modeller so I can obtain a refined 3D structure. But I get the above error, any help.
Thanks, Evans
On Thu, Jul 22, 2004 at 11:08:44AM -0500, Evans Adolfo Tapia wrote: > I'm using modeller lately, and I haven't had any problems while predicting > the tertiary structure of a target using in the multiple alignment > sequences with pdb files. > > Now that I'm trying to use, only CA files I receive this message and > obviously Modeller stops the execution... > > > Array too small. Increase MPRMCNS > > current maximum, current need: 1393200 1393224 > > Do you know if this is a parameter I can increase? or what can be causing > the error.
This error means that Modeller has run out of space for restraint parameters. It guesses the dimension for the restraints array from the number of residues in your system - presumably you have more restraints per residue than it's expecting. You cannot increase this parameter directly, but you can increase Modeller's maximum number of residues to give a correspondingly larger number of restraints - just add
SET MAXRES = 5000
before you read in your PDB files. You may have to tweak this number to use memory most efficiently.
Note also that Modeller's guess depends on your topology model. It allocates memory for roughly 400 restraint parameters per atom, and in turn in the default topology model (3) it assumes roughly 12 atoms per residue. If you're building an all-atom model, then make sure you have SET TOPOLOGY_MODEL = 1 to tell Modeller this - then it'll assume roughly 20 atoms per residue.
Ben Webb, Modeller Caretaker
-
Evans Adolfo Tapia -
Modeller Caretaker