[MMTK] protein setup

Ben Carrington bjc32@hermes.cam.ac.uk
Thu, 15 May 2003 16:01:09 +0100 (BST)

Dear Asbjorn,

As fantastic as MMTK is, I think even Konrad would admit defeat on the
issue of protein folding ;)

The closest you will get to generating a protein structure from the
primary sequence is by some sort of homology modelling / threading (and
even then the target and modelled sequences have to be virtually identical
to get a sensible result!). You said "..WITHOUT reading an experimental
PDB file (of the entire protein).." If you already have some of the
structure and are modelling small gaps, this makes it slightly easier (but
again, not exactly reliable). I think there is the functionality within
MMTK to model small structure gaps - it's a matter of putting together the
right routine. But to answer your question directly, translating a letter
sequence to a sensible all-atom protein model with positions defined, is
just not possible...at this moment in time.

If you are still unsure, feel free to ask.


Ben Carrington
Department of Pharmacology
University of Cambridge, UK

On Thu, 15 May 2003, Asbjorn Christensen wrote:

> Dear MMTK users,
> is there any simple way to initialize the position of atoms
> in a protein directly from a given letter sequence ["Ala", "Gly", "Tyr", ...]
> WITHOUT reading an experimental PDB file (of the entire protein)?
> Of course, this be done in a variety of not non-sense ways, e.g. a
> "linear" chain. If anybody has a simple example of translating
> a letter sequence to a sensible all-atom protein model (with positions defined),
> this will be appriciated.
> PS:  is Examples/protein_construction.py in the online dokumentation
>        up to date? Our MMTK installation has
>        highest "last revision:" ==  2002-6-3
> Best wishes, Asbjorn
> --
> Asbjorn Christensen, PhD
> CAMP, Physics Department, Building 307
> Technical University of Denmark,
> DK-2800 Lyngby, Denmark
> E-mail:       asbjorn@fysik.dtu.dk
> Telephone: (+45) 45 25 32 27
> Telefax:      (+45) 45 93 23 99
> office:         room 257
> _______________________________________________
> mmtk maillist  -  mmtk@starship.python.net
> http://starship.python.net/mailman/listinfo/mmtk