Albert, any news on the issue?
Christoph 2017-10-13 11:02 GMT-06:00 Christoph Junghans <[email protected]>: > Albert, > > 2017-10-13 3:50 GMT-06:00 Albert Solernou <[email protected]>: >> Hi Cristoph, >> I am afraid your answer does not fully address the point. See the comments >> inlined, please. >> >> On Tuesday, October 3, 2017 at 8:34:46 PM UTC+1, Christoph Junghans wrote: >>> >>> 2017-10-03 3:45 GMT-06:00 Albert Solernou <[email protected]>: >>> > Hi Christoph, >>> > if you have a closer look at csgDump_interface, you'll see that masses >>> > are >>> > always fine, but atom types and residue numbers become messed up in the >>> > "List of molecules" section. Line 184 says: >>> > 2 Name 3:GLN:CA Type H Mass 12.01 Resnr 2 Resname GLN >>> You can just do: >>> $ gmx dump -s nvt-100_interface.tpr | less >>> and look for the atoms block, you will find: >>> atom[ 2]={type= 2, typeB= 2, ptype= Atom, m= 1.20100e+01, >>> q=-3.10000e-03, mB= 1.20100e+01, qB=-3.10000e-03, resind= 2, >>> atomnumber= 6} >>> That pretty much matches up with the above line, >> >> >> When I said I don't understand "gmx dump", I meant I do not know how you go >> from here to a PDB line. The error that is showing csg_dump, in the >> previous line is that 3:GLN:CA is not Type H. That is the most important >> thing that I think you should look up. > Ok, I see what you mean! > > VOTCA's identifier scheme is <resnr>:<resname>:<atomname>, so the last > part is the atomname not the atomtype. Atomtypes would make for a > unique identifier string. > > However look again at the output of "gmx dump" I see: > atom[2]={name="CA"} > > So VOTCA basically just picks up the name from inside the tpr, which > is obviously the wrong name, but there isn't much we can do about > that. > >> >>> >>> looking in the >>> "residue" block below: >>> residue[2]={name="GLN", nr=32, ic=' '} >>> So the residue name is correct, too. >> >> >> Yes, the residue name is correct. I always said that it was numbering and >> atom types. >>> >>> >>> The only thing VOTCA does differently is that it start numbering from >>> 1, while gromacs internally starts with 0, hence "3:GLN:CA" (and not >>> "2:GLN:CA"). >>> >>> > a clear example of a wrong type. Then, if you move down to line 708, a >>> > valine residue has the residue number wrong too. >>> > 526 Name 141:VAL:N Type CT Mass 14.01 Resnr 140 Resname VAL >>> Again: >>> atom[ 526]={type= 8, typeB= 8, ptype= Atom, m= 1.40100e+01, >>> q=-4.15700e-01, mB= 1.40100e+01, qB=-4.15700e-01, resind= 140, >>> atomnumber= 7} >>> and >>> residue[140]={name="VAL", nr=73, ic=' '} >>> >>> Do I miss something here? >> >> >> Yes, here you miss that the residue number goes crazy and that every atom in >> this region has belongs to a different VAL, while in reality that is not the >> case. However, I see that this is the numbering that spits "gmx dump", so >> that should not be your bug, if I understood your comment correctly, > Yeah I would ask on the gmx developer list > ([email protected]) about this odd behavior. It seems to me > something in your pre-processing went wrong. > > Christoph > >> >> >> Thanks, >> Albert >>> >>> Christoph >>> >>> > 527 Name 142:VAL:H Type HC Mass 1.008 Resnr 141 Resname VAL >>> > 528 Name 143:VAL:CA Type HC Mass 12.01 Resnr 142 Resname VAL >>> > 529 Name 144:VAL:HA Type C Mass 1.008 Resnr 143 Resname VAL >>> > 530 Name 145:VAL:CB Type O Mass 12.01 Resnr 144 Resname VAL >>> > 531 Name 146:VAL:HB Type N Mass 1.008 Resnr 145 Resname VAL >>> > 532 Name 147:VAL:CG1 Type H Mass 12.01 Resnr 146 Resname VAL >>> > 533 Name 148:VAL:HG11 Type H Mass 1.008 Resnr 147 Resname VAL >>> > 534 Name 149:VAL:HG12 Type C Mass 1.008 Resnr 148 Resname VAL >>> > 535 Name 150:VAL:HG13 Type O Mass 1.008 Resnr 149 Resname VAL >>> > 536 Name 151:VAL:CG2 Type N Mass 12.01 Resnr 150 Resname VAL >>> > 537 Name 152:VAL:HG21 Type H Mass 1.008 Resnr 151 Resname VAL >>> > 538 Name 153:VAL:HG22 Type CT Mass 1.008 Resnr 152 Resname VAL >>> > 539 Name 154:VAL:HG23 Type H1 Mass 1.008 Resnr 153 Resname VAL >>> > 540 Name 155:VAL:C Type CT Mass 12.01 Resnr 154 Resname VAL >>> > 541 Name 156:VAL:O Type HC Mass 16 Resnr 155 Resname VAL >>> > >>> > I will try to reproduce the error with a smaller selection of atoms, >>> > though >>> > I am pretty busy until the 16th Oct. Notes on how to reproduce the error >>> > were attached previously, and should not be complicated to follow. >>> > >>> > I don't know how to interpret the large amount of lines that "gmx dump" >>> > prints out, so to check that its output is correct: >>> > - you could check that with the previously attached .tpr >>> > - I could follow some instructions on how to use "gmx dump". >>> > >>> > >>> > Best, >>> > Albert >>> > >>> > On Sunday, October 1, 2017 at 7:45:10 PM UTC+1, Christoph Junghans >>> > wrote: >>> >> >>> >> Albert, >>> >> >>> >> 2017-10-01 3:20 GMT-06:00 Albert Solernou <[email protected]>: >>> >> > Hi Cristoph, >>> >> > thanks for your reply, and for pointing to the code, and sorry I >>> >> > forgot >>> >> > to >>> >> > attach the messy output. It is now attached, as well as the >>> >> > conflicting >>> >> > .tpr >>> >> > obtained just following the previous notes. >>> >> > >>> >> > It would be great to know what do you think of it. >>> >> Sorry, but what kind of output are you expecting? In your >>> >> "csgDump_interface" atomtypes seem to be consistent with the atom >>> >> names, same for the masses. >>> >> >>> >> And have you checked that "gmx dump" outputs the right number? >>> >> >>> >> Plus, could you make a simple system with just 2 residues exhibit the >>> >> same misbehavior? (That would be easier to us to debug.) >>> >> >>> >> >>> >> Christoph >>> >> > >>> >> > Thanks, >>> >> > Albert >>> >> > >>> >> > >>> >> > >>> >> > On Friday, September 29, 2017 at 9:58:06 PM UTC+1, Christoph Junghans >>> >> > wrote: >>> >> >> >>> >> >> Hi Albert, >>> >> >> >>> >> >> 2017-09-28 6:24 GMT-06:00 Albert Solernou <[email protected]>: >>> >> >> > Hi All, >>> >> >> > I think that there is a bug in csg_dump, as it dumps wrong atom >>> >> >> > types >>> >> >> > and >>> >> >> > some wrong numbers when working with some modified tpr files. >>> >> >> Can you be a bit more explicit which numbers are wrong? >>> >> >> >>> >> >> > >>> >> >> > Not sure how much the bug affects my work, but feel I should >>> >> >> > report >>> >> >> > it >>> >> >> > anyway. I attach clear instructions on how to get the bug, as well >>> >> >> > as >>> >> >> > two >>> >> >> > input files and the wrong output of the csg_dump. I can send the >>> >> >> > whole >>> >> >> > set >>> >> >> > of files but it is 145M. >>> >> >> > >>> >> >> > Could you please have a look at this? >>> >> >> When csg_dump reads tpr files it uses Gromacs' libgmx, so the first >>> >> >> step would be to run "gmx dump" on your tpr files and check if the >>> >> >> same "wrong numbers" show up. >>> >> >> Honestly VOTCA's wrapper around libgmx is very thin, see: >>> >> >> >>> >> >> >>> >> >> >>> >> >> <https://github.com/votca/csg/blob/master/src/libcsg/modules/io/gmxtopologyreader.cc#L93> >>> >> >> >>> >> >> Christoph >>> >> >> > >>> >> >> > Many thanks, >>> >> >> > Albert >>> >> >> > >>> >> >> > -- >>> >> >> > You received this message because you are subscribed to the Google >>> >> >> > Groups >>> >> >> > "votca" group. >>> >> >> > To unsubscribe from this group and stop receiving emails from it, >>> >> >> > send >>> >> >> > an >>> >> >> > email to [email protected]. >>> >> >> > To post to this group, send email to [email protected]. >>> >> >> > Visit this group at https://groups.google.com/group/votca. >>> >> >> > For more options, visit https://groups.google.com/d/optout. >>> >> >> >>> >> >> >>> >> >> >>> >> >> -- >>> >> >> Christoph Junghans >>> >> >> Web: http://www.compphys.de >>> >> > >>> >> > -- >>> >> > You received this message because you are subscribed to the Google >>> >> > Groups >>> >> > "votca" group. >>> >> > To unsubscribe from this group and stop receiving emails from it, >>> >> > send >>> >> > an >>> >> > email to [email protected]. >>> >> > To post to this group, send email to [email protected]. >>> >> > Visit this group at https://groups.google.com/group/votca. >>> >> > For more options, visit https://groups.google.com/d/optout. >>> >> >>> >> >>> >> >>> >> -- >>> >> Christoph Junghans >>> >> Web: http://www.compphys.de >>> > >>> > -- >>> > You received this message because you are subscribed to the Google >>> > Groups >>> > "votca" group. >>> > To unsubscribe from this group and stop receiving emails from it, send >>> > an >>> > email to [email protected]. >>> > To post to this group, send email to [email protected]. >>> > Visit this group at https://groups.google.com/group/votca. >>> > For more options, visit https://groups.google.com/d/optout. >>> >>> >>> >>> -- >>> Christoph Junghans >>> Web: http://www.compphys.de >> >> -- >> You received this message because you are subscribed to the Google Groups >> "votca" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected]. >> To post to this group, send email to [email protected]. >> Visit this group at https://groups.google.com/group/votca. >> For more options, visit https://groups.google.com/d/optout. > > > > -- > Christoph Junghans > Web: http://www.compphys.de -- Christoph Junghans Web: http://www.compphys.de -- You received this message because you are subscribed to the Google Groups "votca" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/votca. For more options, visit https://groups.google.com/d/optout.
