Hi!!

I still have a problem with my modelling of the nucleic acid protein complex... >From the advice of Benn I modified my restyp.lib

... 139 | TP2 | # | TP2 | patch to convert tyrosine to 140 | TP2 | # | TP2A | patch to convert tyrosine to 141 | URA | U | URA |

I understood from the restyp.lib that in my pdb file, the nucleic acid should look like this:

ATOM 3898 P URA B 3 22.300 68.397 94.266 1.00 0.00 ATOM 3899 O1P URA B 3 21.295 67.436 94.787 1.00 0.00 ATOM 3900 O2P URA B 3 21.857 69.524 93.402 1.00 0.00 ATOM 3901 O5' URA B 3 23.422 67.587 93.471 1.00 0.00 ATOM 3902 C5' URA B 3 24.753 68.087 93.372 1.00 0.00 ATOM 3903 C4' URA B 3 25.696 66.981 92.959 1.00 0.00 ATOM 3904 O4' URA B 3 25.472 65.850 93.833 1.00 0.00 ATOM 3905 C3' URA B 3 25.512 66.450 91.538 1.00 0.00 ATOM 3906 O3' URA B 3 26.778 66.021 91.027 1.00 0.00 ATOM 3907 C2' URA B 3 24.612 65.248 91.733 1.00 0.00 ATOM 3908 C1' URA B 3 25.079 64.714 93.075 1.00 0.00 ATOM 3909 N1 URA B 3 23.974 63.957 93.774 1.00 0.00 ATOM 3910 C2 URA B 3 23.611 62.709 93.305 1.00 0.00 ATOM 3911 O2 URA B 3 24.158 62.206 92.317 1.00 0.00 ATOM 3912 N3 URA B 3 22.630 62.041 93.972 1.00 0.00 ATOM 3913 C4 URA B 3 22.002 62.568 95.062 1.00 0.00 ATOM 3914 O4 URA B 3 21.126 61.840 95.609 1.00 0.00 ATOM 3915 C5 URA B 3 22.386 63.855 95.540 1.00 0.00 ATOM 3916 C6 URA B 3 23.369 64.508 94.860 1.00 0.00 ATOM 3917 O2' URA B 3 24.966 64.464 90.615 1.00 0.00 ...

However, when I run modeller all I got is a big, ugly core!! Modeller stops and send to my terminal this message:

******************************* Ludwig@origin 191% mod 1A1Vmov.top

lib-4190 : UNRECOVERABLE library error

Encountered during a sequential formatted READ from an internal file (character variable) IOT Trap /usr/people/ludwig/bin/modeller6v2/bin/mod6v2[38]: 22776 Abort(coredump)

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

I dont think that my problem is related to my .ali or .top script, because if I run the same top script on a modified version of the ali file, without the nucleic acid using the same template pdb file, I still got the core and the same error message. However, if I remove the Nucleic acid coordinates from the pdb file, modeller runs smoothly...

If you were wondering what is in the last lines of the log file when modeller dies:

# Code #_Res #_Segm PDB_code Name ---------------------------------------------------------------------------- --- 1 1CU1 645 1 1CU1 template 2 1A1V 429 1 1A1V template 3 1A1Vm 645 1 1A1Vm model TOP_________> 125 107 CHECK_ALIGNMENT

check_a_343_> >> BEGINNING OF COMMAND openf5__224_> Open 11 OLD SEQUENTIAL ./1CU1.pdb openf5__224_> Open 11 OLD SEQUENTIAL ./1A1V.pdb

Any ideas about where can lie the problem??? Please let me know if you need the files to reproduce the problem.

Thank you for your help and Best regards,

Luis

-----Mensaje original----- De: Modeller Caretaker [mailto:modeller-care@salilab.org] Enviado el: Domingo, 04 de Julio de 2004 03:28 Para: Luis Rosales CC: modeller_usage@salilab.org Asunto: Re: [modeller_usage] Nucleic Acids and modeller

On Fri, Jun 25, 2004 at 05:19:44PM -0500, Luis Rosales wrote: > I need to model a protein complexed to a nucleic acid... > I know my alignment should look like: > PROTEIN/NUCLEIC-ACID > But I was wondering if I should define new libraries in my top file

There shouldn't be any need to do that. If you look in modlib/top_heav.lib (for example - that's for the heavy-atom topology) you should find defined topologies for your standard nucleic acids (i.e. URA, THY, GUA, CYT, ADE residue types).

> and how do you tell modeller: > "Hey, this is a nucleic acid, A stands for adenine!!" (I will be > working with poly Uracil...)

These residues are also listed in modlib/restyp.lib. You simply need to change the 1-letter code for each nucleic acid (currently '#' for every one) to something more useful, and then use the codes you chose there in your alignment file. (Obviously you can't use 'A' for adenine though as it is already used to represent alanine, for example.)

Ben Webb, Modeller Caretaker