Dear Amber user:
I tried to calculate the energy decomposition based on a pairwise per-residue using MM-PBSA, the method developped by Dr. Gohlke in Amber8. I tested the example of Raf-Ras in the program (06_Decomp_pair), it works well and takes about half of hour on my workstation.
My protein system consists of 420 residues (receptor 260 aa, ligand 160 aa). I want to do all residues decomposition, so I set "COMREC=161-420, COMLIG=1-160, and COMPRI=1-420". After finishing the initial energy calculations, the job is stuck on the step "Reading file,....Checking CALC", and after several hours, it showed error "segmentation fault" and aborted.
Since it is probably due to the memory limitation on my workstation (CPU 1.6G, memory 2G), I tried to use a small set of residues, such as "COMREC=161-261, COMLIG=1-160, and COMPRI=1-261". THis time the computation passed the step "Reading file..." and start to process the data, however, it is stuck on the step "Calc delta from raw data" again. It took more than two days and still not finished.
My question is, is it so time-consuming for the pairwise energy decomposisiton? It seems to me a system of 260 residues is not big. I indeed got the results for decomposite only 40 residues quickly (COMREC=20-40, COMLIG=1-20, COMPRI=1-40), it took less than one hour, I wonder why it is so dramatically increased for processing a system of less than 300 residues? Is it normal?
In addition, I am only interested in the energy decompositon between residue pair of receptor and ligand, and the program actually did all pairwise residues including receptor and ligand itself. Can I change some settings to skip such "intra-residue" decomposition? I thought maybe this can speed up the calculation.
Thanks for your advices.
Xin
---
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Mon May 30 2005 - 04:53:02 PDT
