The site is organized as follows:
Due to the increasing amount of sequences available (thanks to the world wide sequencing effort) the gap between the number of available experimental structures of proteins (and more generally of biological macromolecules) is widening despite the improvement of techniques and speed at which experimental structures are solved. By now, the PDB (Protein Data Bank) contains more than 20000 structures which looks unsignificant compared to the millions of available protein sequences known.
This situation might seem catastrophic at first sight but sequences can be gathered in families and it happens frequently that the experimental structure of a member of this family has been solved experimentally. This structure might be used as a template to model the 3D structure of other members of this family and conversely the knowledge of the family can be used to map phylogenic informations onto the known 3D structure. This process of mapping the evolutionnary trace onto a 3D structure can often lead to the identification of zones that play important biological roles (like catalytic sites and partner recognition sites). Finally the knowledge of an experimental structure and an alignment with of the family gives more informations than the structure alone and the alignment alone.
Most of the time however the family of interest has no members whose structure has been solved but even in this difficult case it might be possible to propose an homology model for some members of the family and once done start the previously described 3D-phylogenic analysis. It happens quite often that proteins unrelated at the sequence and functional level have close 3D stuctures. It has even been noticed that the number of novel protein folds discovered each year does not increase significally, which might mean that the number of protein folds is finite. This fact is at the heart of the homolgy modelling approach which consists of three steps: identify proteins whith known structures sharing the same fold of the sequence of interest (process knows as threading), align the sequence of interest with the 3D structure and build the model. In this case, the difficulties are the choice of template structures and the correctness of the alignment because homology modelling sotfwares can not overcome wrong template choices and alignment errors.
We have developped a software ViTO which tries to address these questions and helps the biologist to analyse alignments of proteins (including structures) and prepare the alignments for homology modelling.
As a nice picture is always better than a long talk, here are screenshots that present the components of the program. ViTO is an interactive program coupling a full featured alignment editor with a 3D viewer. The alignment editor and the 3D panel are connected together and each time the alignment is modified the 3D display is updated and conversely. The program has also a macro language which help to produce publication quality images and to add new functionnalities.
The top picture correspond to the alignment panel and the second one to the 3D window with a close-up view on some residues. It is possible that your WEB browser degrades the quality of images.
ViTO has not been published yet but the program is released anyway because we need feedback from users to drive further developpments. The program is functional but there are surely some bugs left that we don't know yet. If you use ViTO and like it please cite it.
V.Catherinot -- CBS BioInformatic Team -- labesse@cbs.cnrs.fr