Hi all,
I am trying to use the soap_score utility to compute the orientational soap protein-protein (pp) score for a set of pp decoys, but I was unable to find a compatible hdf5 file for pp. Is there any way to use the tool for pp?
Thanks
Pablo
On 8/28/17 7:46 AM, Pablo Chacon wrote: > I am trying to use the soap_score utility to compute the orientational > soap protein-protein (pp) score for a set of pp decoys, but I was unable > to find a compatible hdf5 file for pp. > Is there any way to use the tool for pp?
SOAP-PP is not orientation dependent. Its formulation is a little different to the other SOAP potentials (it includes a pairwise term and also an atomistic term) so I'm relatively confident that IMP's soap_score utility won't work with it. There is support in Modeller for SOAP-PP scoring though, and the HDF5 files for SOAP-PP are included in the Modeller distribution. I've updated https://salilab.org/SOAP/ with this information.
Ben
Hi Ben,
Thanks for the replay. I am newie to modeller. Do you have an example where soap_pp.Assessor class is used? (or where I can find it)
Thanks a lot for your help.
Pablo
2017-09-06 19:42 GMT+02:00 Ben Webb ben@salilab.org:
> On 8/28/17 7:46 AM, Pablo Chacon wrote: > >> I am trying to use the soap_score utility to compute the orientational >> soap protein-protein (pp) score for a set of pp decoys, but I was unable to >> find a compatible hdf5 file for pp. >> Is there any way to use the tool for pp? >> > > SOAP-PP is not orientation dependent. Its formulation is a little > different to the other SOAP potentials (it includes a pairwise term and > also an atomistic term) so I'm relatively confident that IMP's soap_score > utility won't work with it. There is support in Modeller for SOAP-PP > scoring though, and the HDF5 files for SOAP-PP are included in the Modeller > distribution. I've updated https://salilab.org/SOAP/ with this > information. > > Ben > -- > ben@salilab.org https://salilab.org/~ben/ > "It is a capital mistake to theorize before one has data." > - Sir Arthur Conan Doyle > _______________________________________________ > IMP-users mailing list > IMP-users@salilab.org > https://salilab.org/mailman/listinfo/imp-users >
On 9/7/17 1:05 AM, Pablo Chacon wrote: > Thanks for the replay. I am newie to modeller. Do you have an example > where soap_pp.Assessor class is used? (or where I can find it)
Just to be absolutely clear here - SOAP-PP contains two scoring terms, a SAS term that is applied to each atom, and an atomistic pairwise statistical potential that is applied to each *pair* of atoms. IMP does have classes to apply both terms (and the potentials themselves are also included with IMP) but the soap_score utility only applies the pairwise term. If you are only ranking rigid docking solutions, the atomistic term should cancel out anyway, so it isn't needed. What do you need to do that isn't handled already by soap_score?
If you want to assess with both terms with Modeller, something like the following should work (although untested):
import modeller import modeller.soap_pp from modeller.scripts import complete_pdb
env = modeller.environ() env.libs.topology.read(file='$(LIB)/top_heav.lib') env.libs.parameters.read(file='$(LIB)/par.lib')
sp = modeller.soap_pp.Asssesor()
mdl = complete_pdb(env, 'my.pdb') atmsel = modeller.selection(mdl) score = atmsel.assess(sp)
Ben
Hi Ben,
You are right, I only need the pairwise term and in principle the soap_score utility is exactly what I need for ranking docking solutions. My confusion was that I thought that I need to include a orientation potential file in order to get the correct soap_pp because it is an orientational dependent potential. I now realize that the default "distance" potential is already orientational.
Thanks for your help and patience,
Pablo
2017-09-07 22:08 GMT+02:00 Ben Webb ben@salilab.org:
> On 9/7/17 1:05 AM, Pablo Chacon wrote: > >> Thanks for the replay. I am newie to modeller. Do you have an example >> where soap_pp.Assessor class is used? (or where I can find it) >> > > Just to be absolutely clear here - SOAP-PP contains two scoring terms, a > SAS term that is applied to each atom, and an atomistic pairwise > statistical potential that is applied to each *pair* of atoms. IMP does > have classes to apply both terms (and the potentials themselves are also > included with IMP) but the soap_score utility only applies the pairwise > term. If you are only ranking rigid docking solutions, the atomistic term > should cancel out anyway, so it isn't needed. What do you need to do that > isn't handled already by soap_score? > > If you want to assess with both terms with Modeller, something like the > following should work (although untested): > > import modeller > import modeller.soap_pp > from modeller.scripts import complete_pdb > > env = modeller.environ() > env.libs.topology.read(file='$(LIB)/top_heav.lib') > env.libs.parameters.read(file='$(LIB)/par.lib') > > sp = modeller.soap_pp.Asssesor() > > mdl = complete_pdb(env, 'my.pdb') > atmsel = modeller.selection(mdl) > score = atmsel.assess(sp) > > > Ben > -- > ben@salilab.org https://salilab.org/~ben/ > "It is a capital mistake to theorize before one has data." > - Sir Arthur Conan Doyle > _______________________________________________ > IMP-users mailing list > IMP-users@salilab.org > https://salilab.org/mailman/listinfo/imp-users >
On 09/08/2017 08:14 AM, Pablo Chacon wrote: > My confusion > was that I thought that I need to include a orientation potential file > in order to get the correct soap_pp because it is an orientational > dependent potential. > I now realize that the default "distance" potential is already > orientational.
No, that's not correct. SOAP-PP is independent of orientation - it simply acts on the distance between pairs of atoms. Some of the SOAP potentials *are* orientation dependent (they act between pairs of pairs of atoms, and take both the distance and some of the internal angles into account) but SOAP-PP is not one of them.
Ben
-
Ben Webb -
Pablo Chacon