rdock-list-def Mailing List for rDock

Dear rDockers,

I found the problem. My SDF records contained the molecules' SMILES 
descriptions as properties and and it seems that for larger molecules 
the string was too long for rDock.
I could fix this by removing the corresponding fields from my SDF and 
now it runs fine.

Cheers,
Jose Manuel

Dear rDock communauty,

I noticed a strange behavior of rDock recently. Here is the short story...

I tried to dock a database on multiple CPUs without using a job 
scheduler, as none was already installed on this particular computer and 
I needed to start the docking quickly for the weekend...

To do so, I followed this tutorial (1). I have 4 CPUs on my computer, so 
I splitted the input ligands file in 4 even parts and I 'submitted' 4 
jobs using the following syntax in a bash script (snippet at reference 2).
When I checked, all jobs were running simultaneously, as I wished.
However, I noticed a very strange behavior when I came back this 
morning. No job finished but all stopped without printing any error 
message at very different times (3). Here is an example of the relevant 
part of one of the log files : (4).

Did anyone try the tutorial (1) and noticed the same behavior? Do you 
have insight about what could have caused it?
Could this be resolved by a job scheduler, such as PBS or Condor?

Thank you for your help!!!

Jose Manuel

Additional Informations :
OS: Centos 6.5
CPUs type :  Intel(R) Core(TM) i7 CPU 950 @ 3.07GH

REFERENCES :
(1) http://rdock.sourceforge.net/how-to-run-rdock-in-parallel/

(2)
#/bin/bash

for file in *sd; do
     out=$(echo $file | cut -d. -f1)'_out'
     rbdock -i $file -o $out -r $protein.prm -p dock.prm -n 20  > 
$out.log 2>&1 &
done

(3)
job1 - Nov 23 11:22
job2 - Nov 20 16:06
job3 - Nov 21 00:29
job4 - Nov 21 10:09

(4)
**************************************************
RECORD #4577
NAME:   previousLigand
RANDOM_NUMBER_SEED:           1164231034

**************************************************
RECORD #4578

Thank you Sergio that solved it!
Jan

On Monday, November 10, 2014, Sergio Ruiz Carmona <sru...@gm...>
wrote:

> Hi Jan,
>
> In the output that rDock writes to the command line you will have
> something like the following text:
>
> ***********************************************
>> The rDock program is licensed under GNU-LGPLv3.0.
>> http://rdock.sourceforge.net/
>> [ ... ]
>> ***********************************************
>>
>> Command line args:
>>  -i FILE.sd
>>  [ ... ]
>>  -n 50
>> Reading polar hydrogens only from ligand SD file
>> DOCKING PROTOCOL:
>> [ ... ]
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> *RbtPharmaSF: Reading mandatory ph4 constraints from FILE.restrReading
>> EXP constraint file(-3,3,-4) 0.5 1.91 0.25 DonCreating ph4 for Dontype
>> EXPReading EXP constraint file(-3,4,-6) 0.5 1.74 0.25 AccCreating ph4 for
>> Acctype EXPReading EXP constraint file(1,1.5,-4) 0.5 1.68 0.25 HydCreating
>> ph4 for Hydtype EXPRbtPharmaSF: No optional ph4 constraints file found*
>> No solvent
>> ------------- Terminate filter 0------------
>> if (SCORE.NRUNS - 49 ) > 0 then
>> 0.0
>> else
>> -1.0
>> end
>>
>> **************************************************
>> RECORD #1
>
> [ ... ]
>
>
> The lines in boldface (from "RbtParmaSF: Reading ..." to "No Solvent")
> correspond to the pharmacophoric constraints: 3 mandatory in "FILE.restr"
> with the corresponding parameters also written and no optional constraints.
> Best regards,
>
> Sergio
>
>
>
> 2014-11-10 12:21 GMT+01:00 Jan Domanski <ja...@gm...
> <javascript:_e(%7B%7D,'cvml','ja...@gm...');>>:
>
>> Hi guys,
>>
>> I have a section pharma defined in my dock.prm file, pointing to some
>> .const files. How to make sure that rdock has read those in correctly and
>> is actually using them?
>>
>> Thanks,
>>
>> - Jan
>>
>>
>> ------------------------------------------------------------------------------
>>
>> _______________________________________________
>> rDock-list-def mailing list
>> rDo...@li...
>> <javascript:_e(%7B%7D,'cvml','rDo...@li...');>
>> https://lists.sourceforge.net/lists/listinfo/rdock-list-def
>>
>>
>

Hi Jan,

In the output that rDock writes to the command line you will have something
like the following text:

***********************************************
> The rDock program is licensed under GNU-LGPLv3.0.
> http://rdock.sourceforge.net/
> [ ... ]
> ***********************************************
>
> Command line args:
>  -i FILE.sd
>  [ ... ]
>  -n 50
> Reading polar hydrogens only from ligand SD file
> DOCKING PROTOCOL:
> [ ... ]
>
>
>
>
>
>
>
>
>
>
> *RbtPharmaSF: Reading mandatory ph4 constraints from FILE.restrReading EXP
> constraint file(-3,3,-4) 0.5 1.91 0.25 DonCreating ph4 for Dontype
> EXPReading EXP constraint file(-3,4,-6) 0.5 1.74 0.25 AccCreating ph4 for
> Acctype EXPReading EXP constraint file(1,1.5,-4) 0.5 1.68 0.25 HydCreating
> ph4 for Hydtype EXPRbtPharmaSF: No optional ph4 constraints file found*
> No solvent
> ------------- Terminate filter 0------------
> if (SCORE.NRUNS - 49 ) > 0 then
> 0.0
> else
> -1.0
> end
>
> **************************************************
> RECORD #1

[ ... ]


The lines in boldface (from "RbtParmaSF: Reading ..." to "No Solvent")
correspond to the pharmacophoric constraints: 3 mandatory in "FILE.restr"
with the corresponding parameters also written and no optional constraints.
Best regards,

Sergio



2014-11-10 12:21 GMT+01:00 Jan Domanski <ja...@gm...>:

> Hi guys,
>
> I have a section pharma defined in my dock.prm file, pointing to some
> .const files. How to make sure that rdock has read those in correctly and
> is actually using them?
>
> Thanks,
>
> - Jan
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> rDock-list-def mailing list
> rDo...@li...
> https://lists.sourceforge.net/lists/listinfo/rdock-list-def
>
>

Hi guys,

I have a section pharma defined in my dock.prm file, pointing to some
.const files. How to make sure that rdock has read those in correctly and
is actually using them?

Thanks,

- Jan

Oh, I see. Many thanks for the answer Daniel!
I'll send any further questions on mdmix to its mailing list ;)

- Jan

On 8 November 2014 01:00, Daniel Álvarez García <alg...@gm...>
wrote:

> The file format is briefly described in page 34 of the manual.
> This is an example of a .const file:
>
>  -3.000   3.000   -4.000 0.50  Don
>  -3.000   4.000   -6.000 0.50  Acc
>   1.000   1.500   -4.000 1.00  Hyd
>
> As you see, it's very simple: center, tolerance and type.
> And the section to include in the prm file is this one
> (supposing al three are mandatory and the file is named mandatory.const):
>
> ################################
> # PHARMACOPHORIC RESTRAINTS
> ################################
> SECTION PHARMA
>     SCORING_FUNCTION RbtPharmaSF
>     WEIGHT 1.0
>     CONSTRAINTS_FILE mandatory.const
> END_SECTION
>
> Regarding the transformation of the density map to
> coordinates as I did for HSP90, please write to pymdmix mailing list,
> (https://sourceforge.net/p/mdmix/mailman/) the topic will better fit
> there
> and will help other users as well ;)
>
> Thanks!
>
> Daniel
>
> El 08/11/2014, a las 01:03, Jan Domanski escribió:
>
> Hi Daniel,
>
> Many thanks for a speedy reply! mdmix looks much better than hand-crafted
> tools I have for that purpose ;)
>
> Regarding the pharmacophore, is there a ready tutorial somewhere showing
> how the .const file should be structured? The only thing I could find in
> the manual was adding the mandatory and optional.const files to the .prm
> file. An example of what you did for HSP90 in the paper would be awesome.
>
> Thanks,
>
> - Jan
>
> On 7 November 2014 23:58, Daniel Álvarez García <alg...@gm...
> > wrote:
>
>> Hi Jan,
>>
>> it is not possible to directly use a density file as a pharmacophore
>> model. You will need to pre-process the grid information
>> to extract the relevant coordinates and write a standard pharmacophore
>> file as described in the documentation.
>>
>> If you need help to extract coordinates from a grid file, take a look at
>> this software: mdmix.sourceforge.net.
>> The python package contains useful objects and functions to deal with
>> grid data. Contact me for further info on this.
>>
>> Daniel
>>
>>
>>
>> El 08/11/2014, a las 00:04, Jan Domanski escribió:
>>
>> > Hi guys,
>> >
>> > Is it possible to use density (.dx) files in the rDock pharmacophore
>> model? Do I understand correctly that now the pharmacophore model is just a
>> bunch of xyz for each pharmacophore type with a weight parameter?
>> >
>> > Thanks,
>> >
>> > - Jan
>> >
>> ------------------------------------------------------------------------------
>> > _______________________________________________
>> > rDock-list-def mailing list
>> > rDo...@li...
>> > https://lists.sourceforge.net/lists/listinfo/rdock-list-def
>>
>>
>
>

The file format is briefly described in page 34 of the manual.
This is an example of a .const file:

 -3.000   3.000   -4.000 0.50  Don
 -3.000   4.000   -6.000 0.50  Acc
  1.000   1.500   -4.000 1.00  Hyd

As you see, it's very simple: center, tolerance and type. 
And the section to include in the prm file is this one
(supposing al three are mandatory and the file is named mandatory.const):

################################
# PHARMACOPHORIC RESTRAINTS
################################
SECTION PHARMA
    SCORING_FUNCTION RbtPharmaSF
    WEIGHT 1.0
    CONSTRAINTS_FILE mandatory.const
END_SECTION

Regarding the transformation of the density map to 
coordinates as I did for HSP90, please write to pymdmix mailing list, 
(https://sourceforge.net/p/mdmix/mailman/) the topic will better fit there 
and will help other users as well ;)

Thanks!

Daniel

El 08/11/2014, a las 01:03, Jan Domanski escribió:

> Hi Daniel, 
> 
> Many thanks for a speedy reply! mdmix looks much better than hand-crafted tools I have for that purpose ;)
> 
> Regarding the pharmacophore, is there a ready tutorial somewhere showing how the .const file should be structured? The only thing I could find in the manual was adding the mandatory and optional.const files to the .prm file. An example of what you did for HSP90 in the paper would be awesome. 
> 
> Thanks,
> 
> - Jan
> 
> On 7 November 2014 23:58, Daniel Álvarez García <alg...@gm...> wrote:
> Hi Jan,
> 
> it is not possible to directly use a density file as a pharmacophore model. You will need to pre-process the grid information
> to extract the relevant coordinates and write a standard pharmacophore file as described in the documentation.
> 
> If you need help to extract coordinates from a grid file, take a look at this software: mdmix.sourceforge.net.
> The python package contains useful objects and functions to deal with grid data. Contact me for further info on this.
> 
> Daniel
> 
> 
> 
> El 08/11/2014, a las 00:04, Jan Domanski escribió:
> 
> > Hi guys,
> >
> > Is it possible to use density (.dx) files in the rDock pharmacophore model? Do I understand correctly that now the pharmacophore model is just a bunch of xyz for each pharmacophore type with a weight parameter?
> >
> > Thanks,
> >
> > - Jan
> > ------------------------------------------------------------------------------
> > _______________________________________________
> > rDock-list-def mailing list
> > rDo...@li...
> > https://lists.sourceforge.net/lists/listinfo/rdock-list-def
> 
> 

Hi Daniel,

Many thanks for a speedy reply! mdmix looks much better than hand-crafted
tools I have for that purpose ;)

Regarding the pharmacophore, is there a ready tutorial somewhere showing
how the .const file should be structured? The only thing I could find in
the manual was adding the mandatory and optional.const files to the .prm
file. An example of what you did for HSP90 in the paper would be awesome.

Thanks,

- Jan

On 7 November 2014 23:58, Daniel Álvarez García <alg...@gm...>
wrote:

> Hi Jan,
>
> it is not possible to directly use a density file as a pharmacophore
> model. You will need to pre-process the grid information
> to extract the relevant coordinates and write a standard pharmacophore
> file as described in the documentation.
>
> If you need help to extract coordinates from a grid file, take a look at
> this software: mdmix.sourceforge.net.
> The python package contains useful objects and functions to deal with grid
> data. Contact me for further info on this.
>
> Daniel
>
>
>
> El 08/11/2014, a las 00:04, Jan Domanski escribió:
>
> > Hi guys,
> >
> > Is it possible to use density (.dx) files in the rDock pharmacophore
> model? Do I understand correctly that now the pharmacophore model is just a
> bunch of xyz for each pharmacophore type with a weight parameter?
> >
> > Thanks,
> >
> > - Jan
> >
> ------------------------------------------------------------------------------
> > _______________________________________________
> > rDock-list-def mailing list
> > rDo...@li...
> > https://lists.sourceforge.net/lists/listinfo/rdock-list-def
>
>

Hi Jan,

it is not possible to directly use a density file as a pharmacophore model. You will need to pre-process the grid information
to extract the relevant coordinates and write a standard pharmacophore file as described in the documentation. 

If you need help to extract coordinates from a grid file, take a look at this software: mdmix.sourceforge.net.
The python package contains useful objects and functions to deal with grid data. Contact me for further info on this.

Daniel



El 08/11/2014, a las 00:04, Jan Domanski escribió:

> Hi guys, 
> 
> Is it possible to use density (.dx) files in the rDock pharmacophore model? Do I understand correctly that now the pharmacophore model is just a bunch of xyz for each pharmacophore type with a weight parameter?
> 
> Thanks,
> 
> - Jan
> ------------------------------------------------------------------------------
> _______________________________________________
> rDock-list-def mailing list
> rDo...@li...
> https://lists.sourceforge.net/lists/listinfo/rdock-list-def

Hi guys,

Is it possible to use density (.dx) files in the rDock pharmacophore model?
Do I understand correctly that now the pharmacophore model is just a bunch
of xyz for each pharmacophore type with a weight parameter?

Thanks,

- Jan