What is SPROUTS ?

Historically, this database was designed to gather all the results from a study concerning the comparison between tools devoted to the prediction of stability changes upon point mutations. The amount of data to collect from various sources was important and the need of a database was evident.

10 structures corresponding to a total of 1211 amino acids have been processed each time by 5 different programs for the 19 possible mutations on each amino acid. Thus, the output data at the end of the experiment consisted in 138054 pieces of data. The execution of the different programs is producing one file per amino acid for a total of more than 7200 files to manipulate. It is clear that this solution is not conceivable for an open and easy access. The need of a database is mandatory and also offers the opportunity to provide this information for the whole scientific community.

As of today, this database has grown up and consists in more than 100 structures which have been computed for a total of around 16500 amino acids. The second aim of this database is to offer simple and user-friendly tools to better visualize and analyze the results obtained. We are now able to propose three ways of visualization and analysis: the first one consists in getting raw ΔΔG values in a table. The second one is a 2D graph representation of a computed stability score for each residue of a given sequence and for each tool. The last one is based on a Jmol applet [Jmol] with the possibility to represent the 3D structure of a given protein with symbols representing the information stored in the database. We assume that each visualization mode offers a different look on the data stored in the database and will suit to every scientists willing to query the database whether they are more used to handle 3D protein structure or 1D/2D sequence problems.

Finally, the ultimate objective is to integrate these data and their analyses with other structural bioinformatic concepts in order to improve other methods that may be related to this concept. We are currently working at adding the information extracted from our other projects related to the prediction of protein folding nucleus in order to obtain a meta server devoted to the characterisation of the folding core of proteins.