Centre for Molecular and Biomolecular Informatics Radboud University, Toernooiveld 1, P.O. Box 9010, 6500 GL Nijmegen, +31 (0)24-3653391, postmaster@cmbi.ru.nl

Protein structure

The CMBI is heavily committed to protein structure research. The main research topics are:

1. explaining biology and medicine from structure data
2. homology modelling
3. structure validation
4. databases
5. structure analysis servers
6. low complexity sequence-motif database

We normally do the first of these in collaboration with experimentalists.
Homology modelling helps to get 3D structures to help us understand biology (X-ray and NMR techniques are too slow for us!).
Structure validation is needed to do the first two topics as well as possible.
Databases and information systems are needed to do topics 1-3 as well as possible.
The CMBI is also a service institute, so we make the databases publicly available.

Protein structure

There is no recipe for explaining biology and medicine. The only things that are certain are that you need to know a lot about many different topics and combine that knowledge with as much data as you can find.
On the 'small projects and services' page you can find a small collection of the kind of projects we work on.
In many projects we try to arrive at coordinates via homology modelling. And that bring us to our second hobby:

Homology modelling

We design homology modelling and related software, and we use this software to answers questions from every day practice.
The software design centers around the WHAT IF software.
We give every year a homology modelling course. This course is open for everybody who wants to learn everything about homology modelling. One step in this course consists of a series of articles that are relevant to this topic.
Every two years the CASP competition is held. This competition is used as a benchmark for the quality of homology modelling, threading and secondary structure prediction software. The idea is to compare software from different authors, but also to compare today's state of the art with the state of the art from 2, 4, ... years ago. We have participated every time with the WHAT IF software, and we don't mind at all if you look in the special volumes of proteins for the results.
A WHAT IF based homology modelling server is aprt of our 3D server suite.

Structure validation

Homology modelling requires a template structure. Template structures can be found in the PDB. These structures have been determined experimentally by X-ray or NMR, and consequently they contain errors. We have written software that determines a little more than 100 kinds of errors in macromolecular structures. This software is available as part of the biotech partners validation server, it sits in our 3D server suite.

Additionally, we run the validation software weekly on all new PDB files. The resulting database of validation reports is called PDBREPORT and is available for inspection.

Databases

The three projects mentioned above all rely heavily on data, access to data and combination of data from different sources. We are therefore involved in a large series of database projects.

• DSSP. The DSSP database consists of one DSSP file per PDB file. A DSSP file describes the backbone hydrogenbonds and the secondary structure of the proteins in the PDB files.
• PDBFINDER. The PDBFINDER database is one giant file that holds per PDB file one entry with the most important statistics about that file.
• GPCRDB. One of our hobbies is the study of G protein-coupled receptors (GPCRs). The GPCRDB is not a database but an information system aimed at collecting and disecting GPCR related data.
• PDBSELECT. This is a set of lists of proteins that were at one time considered a representative sub-set from the PDB.

The program WHAT IF is a versatile program that offers thousands of options to analyse protein structures, predict point mutations, calculate macromolecular electrostatics, visualise molecular dynamics movies, analyse cavities, etc. Often this program is needed for one research project and it seems a bit an overkill to install the whole program (and learn how to operate it) if just one picture is needed, or one set of atomic contacts needs to be calculated to convince a referee.
For these cases we made the WHAT IF based servers. Usage of these servers is free, but please do not automate this usage. These servers are meant for individual calculations.
If a new server is needed, please feel free to ask Gert Vriend to build it.