electrostatics / apbs-pdb2pqr Goto Github PK

Note that some of this is done already in https://github.com/Electrostatics/apbs-pdb2pqr/tree/master/sandbox/baker/apbs-python

@lizutah

Send me an email, word doc, or edit the MD files to fill in content on this page:
http://electrostatics.github.io/apbs-pdb2pqr/docs/structures-ready/

Test on pKa Cooperative datasets

There should be a parameter to cmake to control inclusion of the geometric flow code. Something along the lines of ENABLE_GEOFLOW=ON, by default. Thus allowing the user to set it to OFF should they desire.

@lizutah had several suggestions about new, more compact, formats for output of the electrostatic potential from APBS.

In addition to reducing the output size, we need a multilevel output format which could include information about electrostatic potentials at various resolutions. There is already a data structure available for this in APBS (Vmgrid) but no output format. This would be useful for providing output at various stages of a focusing calculation. In particular, it would allow us to write out multi-resolution dielectric and ionic maps which could be read back in for subsequent calculations.

This data format could, perhaps, be an XML-based generalization of OpenDX with individual OpenDX data files included as CDATA fields. This is a kludge but would make backwards compatibility more straightforward.

This is a modified version of JIRA issue APBS-7.

Some users have experienced an apparent memory allocation error when using mg. https://groups.google.com/forum/#!topic/apbs-users/0ygcYogGmuU may be such an example, although I don't actually know the exact circumstances.

In any case, the error is proximate to the problem — it's caused by an integer overflow in one of the array indices. Nothing new here. Unfortunately.

@keith923 and @astevens-pnnl

Add content to this page:

http://electrostatics.github.io/apbs-pdb2pqr/docs/reading-list/

Internal link structure is incorrect.

When disabling BEM an error results in routines.c because the BEM code is not conditionally included.

Send me an email, word doc, or edit the MD file in order to fill in content for this page:

http://electrostatics.github.io/apbs-pdb2pqr/docs/calculating/

On Wed, Apr 2, 2014 at 11:09 AM, Che Martin [email protected] wrote:

Good afternoon,

I am a new user. When I try to generate a pqr file from my attached pbd which contains a phosphorylated Thr (Res 82). The resulting pqr does not contain my phosphorylated residue. I read the mailing list and they mentioned that I should use a modified force field. So I added the phosphate group to my FF file (attached) however the pqr file still comes back missing the modified Thr82 residue and contains a remarked section:

Please advise, am I using the wrong force-field file?


Thank you,
Che
*****************************************************************************************

REMARK   5 WARNING: PDB2PQR was unable to assign charges
REMARK   5          to the following atoms (omitted below):
REMARK   5              63 O in THR 18
REMARK   5              152 O in THR 24
REMARK   5              630 O in THR 56
REMARK   5              716 O in THR 63
REMARK   5              819 O in THR 71
REMARK   5              992 P1 in TPO 82
REMARK   5              993 O1 in TPO 82
REMARK   5              994 O2 in TPO 82
REMARK   5              995 O3 in TPO 82
REMARK   5              1244 O in THR 98
REMARK   5 This is usually due to the fact that this residue is not
REMARK   5 an amino acid or nucleic acid; or, there are no parameters
REMARK   5 available for the specific protonation state of this
REMARK   5 residue in the selected forcefield.
REMARK   5
REMARK   5 WARNING: Non-integral net charges were found in
REMARK   5          the following residues:
REMARK   5              THR A 18 - Residue Charge: 0.5100
REMARK   5              THR A 24 - Residue Charge: 0.5100
REMARK   5              THR A 56 - Residue Charge: 0.5100
REMARK   5              THR A 63 - Residue Charge: 0.5100
REMARK   5              THR A 71 - Residue Charge: 0.5100
REMARK   5              TPO A 82 - Residue Charge: -0.4300
REMARK   5              THR A 98 - Residue Charge: 0.5100

**********************************************************************************************

From: ***
Subject: Re: [apbs-users] calculating the net dipole using APBS

Hi,

Thank you for the quick reply.
I would appreciate it a lot if the dipole information would be available on the next releases of APBS package (quadrupole is least significant).
It is something that would be very useful for me :)

Thanks again!
B***k.

On Tuesday, July 30, 2013 4:50:47 AM UTC+3, Baker, Nathan wrote:
Hi –

We haven’t added that feature yet, partly because it is not obvious whether the Coulombic or full Poisson-Boltzmann multipole moments would be most useful. The Coulombic multipole moments are easy to calculate but the PB multipole moments would require more work.

Thanks,

Nathan Baker
Pacific Northwest National Laboratory
Tel: +1-509-375-3997
http://nabaker.me

From: ***
Subject: Re: [apbs-users] calculating the net dipole using APBS

Hi,

I've encountered this post when searching for a solution for getting the dipole and quadruple moment of a protein.
Have you already integrated this feature in the latest APBS version? If not, how can I get this data?

Thanks,
B***k.

On Wednesday, February 13, 2013 6:06:41 AM UTC+2, Baker, Nathan wrote:
Hello --

We do not currently have an option for reporting the dipole moment of a protein in APBS. This information is actually calculated (thanks to Mike Schnieders) when you run an APBS calculation with the "bcfl sdh" boundaries conditions; however, we don't report the result to the user. It would be relatively straightforward to add print statements to the code (http://sourceforge.net/p/apbs/code/ci/d8b8ef203030c0f4cfc8d98a58868f38240a490c/tree/src/mg/vpmg.c#l2662) to report the monopole, dipole, and quadrupole moments — if that would be of general interest to the APBS community.

Do you feel comfortable modifying the code yourself or would you like us to do that?

Thanks,

Nathan Baker
Pacific Northwest National Laboratory
Phone: +1-509-375-3997

Hi APBS group (Hi Nathan)

I have used APBS to calculate and then visualise the electrostatic potential and electric field lines of a peptide (42 residues) and a ion channel protein. My main interest is in looking at the interaction of the peptide with the ion channel, in particular the orientation of the toxin through the interaction with the electric field of the ion channel. For that it would be good to calculate (and visualise) the net dipole of the peptide. Is ther such an option in APBS?

Since the net dipole is "simply" a function of all the point charges in the system which are probably calculated by APBS for visualising the electrostatic surface potential I thought there might be an option to get the vector for the net dipole of the molecule.
If this is not possible in the current code, any suggestions on how I could calculate the vector for the net dipole of a molecule?

thanks a lot
E*****e

Support for integrating APBS with TINKER (http://dasher.wustl.edu/tinker/) has been broken during the migration from Automake to Cmake to Scons to Cmake to... We should fix it.

A known list of pdb files will not work with PROPKA 3.0 or 3.1. These will will work with PROPKA 1.0. This is probably due to the fallback mechanism not being implemented properly for unknown residues.

@kmonson TODO: Dig up files from email and attach to this ticket.

This may no longer be relevant -- have we decided on Cmake for APBS @keith923 and @kmonson ?

@kmonson

The code is from Dimitri Morikis' group and can be downloaded from http://biomodel.engr.ucr.edu/aesop-1.0.0.zip

Send me an email, word doc, or edit the MD files to fill in content on this page:
http://electrostatics.github.io/apbs-pdb2pqr/docs/structures-ready/

Robert Konecny discovered the following issue on Ubuntu 13.10 (gcc 4.8.1):

>...
>[ 76%] Building CXX object
>src/geoflow/CMakeFiles/apbs_geoflow.dir/surfconcz.cpp.o
>In file included from
>/tmp/SandBox-2014.2.10/apbs-pdb2pqr/apbs/src/geoflow/modules.h:59:0,
>                 from
>/tmp/SandBox-2014.2.10/apbs-pdb2pqr/apbs/src/geoflow/surfconcz.cpp:55:
>/tmp/SandBox-2014.2.10/apbs-pdb2pqr/apbs/src/geoflow/Mat.h: In
>instantiation of ‘T Stencil<T>::deriv(T) const [with T = double]’:
>/tmp/SandBox-2014.2.10/apbs-pdb2pqr/apbs/src/geoflow/surfconcz.cpp:120:57:
>   required from here
>/tmp/SandBox-2014.2.10/apbs-pdb2pqr/apbs/src/geoflow/Mat.h:223:47: error:
>‘dot’ was not declared in this scope, and no declarations were found by
>argument-dependent lookup at the point of instantiation [-fpermissive]
>         double gram = 1.0 + dot(dx(),dy(),dz());
>                                               ^
>In file included from
>/tmp/SandBox-2014.2.10/apbs-pdb2pqr/apbs/src/geoflow/surfconcz.cpp:55:0:
>/tmp/SandBox-2014.2.10/apbs-pdb2pqr/apbs/src/geoflow/modules.h:86:8:
>note:
>‘double dot(double, double, double)’ declared here, later in the
>translation unit
> double dot(double x, double y, double z);
>        ^
>make[2]: *** [src/geoflow/CMakeFiles/apbs_geoflow.dir/surfconcz.cpp.o]
>Error 1
>make[1]: *** [src/geoflow/CMakeFiles/apbs_geoflow.dir/all] Error 2
>make: *** [all] Error 2

Geoflow development was done using gcc 4.4.7-4, so clearly weare doing something that was once copacetic, and is no longer. The temporary 'fix' is to set the -fpermissive flag in the cmake file, as below. This however is not a good permanant solution.

diff --git a/apbs/CMakeLists.txt b/apbs/CMakeLists.txt
index f6fb030..38715a6 100644
--- a/apbs/CMakeLists.txt
+++ b/apbs/CMakeLists.txt
@@ -54,6 +54,7 @@ OPTION (BUILD_SHARED_LIBS "Build shared libraries." OFF)
  message(STATUS "Setting project paths")

  set(CMAKE_C_FLAGS -std=c99)
+set(CMAKE_CXX_FLAGS -fpermissive)

  set(APBS_ROOT ${PROJECT_SOURCE_DIR})
  set(EXECUTABLE_OUTPUT_PATH ${APBS_ROOT}/bin)

Rather than checking for Fortran and printing a sensible error informing the user of the deficiency, we currently bug out sort of implying that it might be maloc.

Currently pdb2pqr only processes one of the models in a pdb file. We need to fix this on a high level by allowing ourselves to select a model when we build the protein object and run them all in a loop.

@keith923 or @lizutah

If a user has loaded all of the necessary maps (dielectric, charge, kappa), then a molecule should not be required. However, the current version of APBS requires this.

This problem was reported by Lane Votapka.

The following tests are not yet in tast_cases.cfg:

• ../examples/born/apbs-mol-fem.in
• ../examples/born/apbs-smol-fem.in
• ../examples/born/apbs-mol-fem-extmesh.in

I think it's a strange shell script right now... will this change if we move to scons?

Test Name: FKBP
Test Directory: ../examples/FKBP
Testing 1d7h-dmso-mol.in
FAILED (1.656350000000e+01; expected 1.500770000000e+01; 10.3667% error)

This depends on the availability of the functionality in the main code...

This may no longer be relevant -- have we decided on Cmake for APBS @keith923 and @kmonson ?

@lizutah

There are various and sundry issues that need dealing with in CMakeLists.txt. It'll be faster and less confusing to document them here, in a single issue.

Ira to use new integrated registration form in new version of website: https://docs.google.com/forms/d/1CsftV09vLGIxeMHwevGy8SDVYKoihs8EWLNjsbjxIRw/viewform

Improved placement of surface points using ITK's new particle filter.

The tool mergedx2 (tools/mesh/mergedx2.c) will not compile on windows under visual studio because it relies on optargs to parse command line arguments.

Remove any dependency on optargs from the code, and find a platform agnostic way to parse the arguments.

@sobolevnrm and @lizutah

Fill in content in three files:

Getting structures ready (structures-ready)
Calculating Electrostatics (calculating)
Visualizing Results (visualizing-results)

files located in docs directory.

Test Name: geoflow
Test Directory: ../examples/geoflow
Testing geoflow.in
PASSED within error (-1.025630000000e+01; expected -1.030410000000e+01; -0.463893% error)
FAILED (5.684660000000e-01; expected 5.417420000000e-01; 4.93298% error)

Recode APBS to use GAMer (http://fetk.org/codes/gamer/ - ask @sobolevnrm for details) or Cleaver (http://www.sci.utah.edu/software/cleaver.html - ask @lizutah for details).

This is a modified version of issue APBS-2 from JIRA.

Easy fix in main.py and main_cgi.py

example.dat is missing from apbs-pdb2pqr/pdb2pqr/examples/1a1p

One of our users alerted us to the following problem, and resolution (yay!).

Setting a membrane potential stopped working some time between apbs
1.3 (compiled on Nov 11 2010 at 14:30:08) and apbs 1.4 (compiled on
Nov 2 2012 at 09:21:27).

Vpbe_getzmem(...) in src/generic/vpbe.c results in:
VASSERT: ASSERTION FAILURE! filename
/Users/d3x874/apbs/src/generic/vpbe.c, line 200, (thee->param2Flag)

Lines 459 to 460 of src/generic/vpbe.c were commented out at some point:
// if (V != VNULL) thee->param2Flag = 1;
// else thee->param2Flag = 0;

It appears that thee->param2Flag is not modified anywhere anymore
(except at allocation of thee). Thus Vpbe_getzmem() and associated
functions will always assert.

Everything worked again after replacing the commented out code with:
if (V != 0.0) thee->param2Flag = 1;
else thee->param2Flag = 0;

I imagine the code was commented out due to compiler warnings about
type mismatch.