QM/MM error for protein with more than 2 chains


Just Got Here
Hi everyone,

when I perform a QM/MM calculation for a protein with more than 2 peptide chains, I get the following error from the pre_short module:

**********
* 0: pre_shorts:increase dimension 2
**********

I removed all the non-standard residues from the pdb file already and successfully calculated a QM/MM energy with just the first one or two chains of the protein included. I also tried this for two other proteins with the same results. Thus, my pdb format seems to be fine, as well as my handling of the termination of one peptide chain (TER keyword).

I looked at the pre_short source code (https://svn.pnl.gov/svn/nwchem/trunk/src/prepar/pre_short.F), which lets me believe that this has something to do with searching for the smallest distance between some atoms (I don't know which atoms and what for). The error message arises from the line:

if(number.gt.10) call md_abort('pre_shorts:increase dimension',2)

Does someone know why this happens in my case and what this "number" variable does? How can I "increase the dimension", as it is suggested? It does not seem like this value 10 can be increased in any way. As I said, this problem only arises when my proteins starts having more than 2 separate chains.

Attaching my prepared pdb and input file somehow did not work. You can take a look at them here:

Just read input: https://pastebin.com/eeUdEy5g
Download pdb and input: https://files.fm/u/hcxxssqq#_

Hopefully someone can help me with this problem.

Best regards,
Cbru94

Just Got Here
Seems to be just an issue of system size
Additional information: I tried it out for different systems now, for which NWChem can handle up to 5 (smaller) chains, but not more. Therefore, it seems to be an issue of the size of the system.

Is there a maximum number of atoms that can be handled with this QM/MM code? It would be quite low then (around 1000). I can't believe this is the reason.

Does anyone know something about this?

Just Got Here
Cbru94,

Due to security reasons, we cannot download your pdb and input files. Is your protein system available in the Protein Data Bank? How many atoms do each of the chains have? Is the pdb more than 100,000 atoms?

-Amity

Just Got Here
Thank you for your reply,

the protein system's ID is 5GS5 (the one I tried to send you, I tried it on more than one system). However, I deleted the heteroatoms and changed the format to fit the one, that NWChem uses, because otherwise it does not work at all.

That system now has approx. 25000 atoms total with 8 chains (ca. 3000 atoms each). Different combinations of 2 or 3 of those chains each work perfectly fine for me.

I have now also tried to delete all the atoms far away from one center point, which results in a system with for example 5 or 6 chains, but with only around 30 atoms for each chain. This also does not work, while the version with just 2 or 3 chains does.

So I am pretty certain now that the above stated error arises from the fact that I have more than 3 or 4 chains (depends on system), regardless of the size of the chains themselves.

How would you suggest, that I can provide you with the input file and pdb file instead? Don't you maybe have an example of a calculation of a large protein system? I mean, considering a whole protein not rarely results in having to consider more than 3 or 4 chains, right?

I would be really grateful for your assistance. Maybe you can just explain to me, what that part of the source code does?

Best,
Cbru94


Forum >> NWChem's corner >> QMMM