[gmx-developers] pdb2gmx: New features?
boristim at mail.ru
Thu Jun 11 15:03:59 CEST 2020
>Четверг, 11 июня 2020, 14:14 +03:00 от Berk Hess <hess at kth.se>:
>For point 1, yes I agree circular molecule generation would be useful. This is not difficult, just a matter of not using termini and looking up the neighbor residue indices periodically. I have needed this once and did not have/take the time to implement this in pdb2gmx. Are you sufficiently good a programming that you could attempt to add this feature yourself?
Yes. I’m ready to try to implement this feature. The code seems clear to me. If any questions arise, may I ask them in private correspondence?
>For point 2, is it sufficient to add the name to residuetypes.dat? Using the concept of a chain sounds nice in principle, but I don't know if this is used consistently in pdb.
I consider this item a little more broadly. Imagine RNA-DNA duplex.
Everything is fine until you make just one replacement of ribonucleotide/deoxyribonucleotide with the opposite, (DNA < - > RNA) in any of the chains. pdb2gmx will produce a chain heterogeneity error, and by editing residuetypes.dat we will eventually (in terms of grromacs) get a RNA/RNA or DNA/DNA duplex model with a very strange residuetypes.dat.
Residuetypes.dat ceases to be a global database and becomes a private project file.
Perhaps it is ideologically more correct to allow the creation of heterogeneous chains, using, for example, the mechanism of index files to create a molecule? As far as I understand, after the topology is created, heterogeneity in the simulation does not matter.
>On 2020-06-10 23:58, Boris Timofeev wrote:
>>Dear gromacs developers!
>>I want to discuss the actual questions.
>>Do we need to adjust/redefine some the pdb2gmx paradigms?
>>1. Unfortunately, pdb2gmx does not know how to work with ring/circular molecules.
>>The terminals must be defined as for proteins, and for dna/rna.
>>At the same time, interest to circular structures found quite a lot, growing, and it is quite difficult to make such a topology manually,
>>because this operation involves atoms removing , renumbering, corrections of charges.
>>Such work will not take much time and effort, but will be very useful in the future.
>>Examples of ring structures 5zo6.pdb, 2oiu.pdb, 2kjf.pdb
>>2. The concept of a "group" of "Protein", "DNA", "RNA", being conditional, greatly complicates the construction of
>>complex molecules containing "non-standard" elements in their chain.
>>On the one hand, a variety of terminal molecules, such as acetic acid, NH2, etc. are included in the group of "Protein" (proteins are in no way), on the other hand, an attempt to include even a protein molecule
>>described in the external rtp-file, but not included in residuetypes.dat, into the protein chain,causes an error.
>>The modification in the DNA sequence of one nucleotide into the RNA, also causes the error, although oligonucleotide chains of heterogeneous elements are now widely used.
>>Attempts to modify oligonucleotide terminals also lead to the need to include the compounds in to the group "DNA"/"RNA"
>>(depending on the task) , which are not related to them, for example, fatty acids.
>>Maybe it makes sense to change the error on a warning about the heterogeneity of the chain
>>and include the concept of "CHAIN" into the group?
>>It is very interesting to know your opinions and discuss any interface details.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the gromacs.org_gmx-developers