Christophe Geourjon

Improved performance in protein secondary structure prediction by inhomogeneous score combination.

MOTIVATIONnIn many fields of pattern recognition, combination has proved efficient to increase the generalization performance of individual prediction methods. Numerous systems have been developed for protein secondary structure prediction, based on different principles. Finding better ensemble methods for this task may thus become crucial. Furthermore, efforts need to be made to help the biologist in the post-processing of the outputs.nnnRESULTSnAn ensemble method has been designed to post-process the outputs of discriminant models, in order to obtain an improvement in prediction accuracy while generating class posterior probability estimates. Experimental results establish that it can increase the recognition rate of protein secondary structure prediction methods that provide inhomogeneous scores, even though their individual prediction successes are largely different. This combination thus constitutes a help for the biologist, who can use it confidently on top of any set of prediction methods. Moreover, the resulting estimates can be used in various ways, for instance to determine which areas in the sequence are predicted with a given level of reliability.nnnAVAILABILITYnThe prediction is freely available over the Internet on the Network Protein Sequence Analysis (NPS@) WWW server at http://pbil.ibcp.fr/NPSA/npsa_server.ht ml. The source code of the combiner can be obtained on request for academic use.

euHCVdb: the European hepatitis C virus database

The hepatitis C virus (HCV) genome shows remarkable sequence variability, leading to the classification of at least six major genotypes, numerous subtypes and a myriad of quasispecies within a given host. A database allowing researchers to investigate the genetic and structural variability of all available HCV sequences is an essential tool for studies on the molecular virology and pathogenesis of hepatitis C as well as drug design and vaccine development. We describe here the European Hepatitis C Virus Database (euHCVdb, ), a collection of computer-annotated sequences based on reference genomes. The annotations include genome mapping of sequences, use of recommended nomenclature, subtyping as well as three-dimensional (3D) molecular models of proteins. A WWW interface has been developed to facilitate database searches and the export of data for sequence and structure analyses. As part of an international collaborative effort with the US and Japanese databases, the European HCV Database (euHCVdb) is mainly dedicated to HCV protein sequences, 3D structures and functional analyses.

Integrated databanks access and sequence/structure analysis services at the PBIL

The World Wide Web server of the PBIL (Pôle Bioinformatique Lyonnais) provides on-line access to sequence databanks and to many tools of nucleic acid and protein sequence analyses. This server allows to query nucleotide sequence banks in the EMBL and GenBank formats and protein sequence banks in the SWISS-PROT and PIR formats. The query engine on which our data bank access is based is the ACNUC system. It allows the possibility to build complex queries to access functional zones of biological interest and to retrieve large sequence sets. Of special interest are the unique features provided by this system to query the data banks of gene families developed at the PBIL. The server also provides access to a wide range of sequence analysis methods: similarity search programs, multiple alignments, protein structure prediction and multivariate statistics. An originality of this server is the integration of these two aspects: sequence retrieval and sequence analysis. Indeed, thanks to the introduction of re-usable lists, it is possible to perform treatments on large sets of data. The PBIL server can be reached at: http://pbil.univ-lyon1.fr.

MPSA: integrated system for multiple protein sequence analysis with client/server capabilities

UNLABELLEDnMPSA is a stand-alone software intended to protein sequence analysis with a high integration level and Web clients/server capabilities. It provides many methods and tools, which are integrated into an interactive graphical user interface. It is available for most Unix/Linux and non-Unix systems. MPSA is able to connect to a Web server (e.g. http://pbil.ibcp.fr/NPSA) in order to perform large-scale sequence comparison on up-to-date databanks.nnnAVAILABILITYnFree to academic http://www.ibcp.fr/mpsa/nnnCONTACTnc.blanchet@ibcp.fr

GeneFarm, structural and functional annotation of Arabidopsis gene and protein families by a network of experts

Genomic projects heavily depend on genome annotations and are limited by the current deficiencies in the published predictions of gene structure and function. It follows that, improved annotation will allow better data mining of genomes, and more secure planning and design of experiments. The purpose of the GeneFarm project is to obtain homogeneous, reliable, documented and traceable annotations for Arabidopsis nuclear genes and gene products, and to enter them into an added-value database. This re-annotation project is being performed exhaustively on every member of each gene family. Performing a family-wide annotation makes the task easier and more efficient than a gene-by-gene approach since many features obtained for one gene can be extrapolated to some or all the other genes of a family. A complete annotation procedure based on the most efficient prediction tools available is being used by 16 partner laboratories, each contributing annotated families from its field of expertise. A database, named GeneFarm, and an associated user-friendly interface to query the annotations have been developed. More than 3000 genes distributed over 300 families have been annotated and are available at http://genoplante-info.infobiogen.fr/Genefarm/. Furthermore, collaboration with the Swiss Institute of Bioinformatics is underway to integrate the GeneFarm data into the protein knowledgebase Swiss-Prot.

Conservation of amino acids into multiple alignments involved in pairwise interactions in three-dimensional protein structures.

We present an original strategy, that involves a bioinformatic software structure, in order to perform an exhaustive and objective statistical analysis of three-dimensional structures of proteins. We establish the relationship between multiple sequences alignments and various structural features of proteins. We show that amino acids implied in disulfide bonds, salt bridges and hydrophobic interactions have been studied. Furthermore, we point out that the more variable the sequences within a multiple alignment, the more informative the multiple alignment. The results support multiple alignments usefulness for predictions of structural features.

SuMo: A Tool for Protein Function Inference Based on 3D Structures Comparisons

The prediction of important residues for binding/recognition sites in protein 3D structures is still a matter of challenge. Indeed, binding sites recognition is generally based on geometry often combined with physico-chemical properties of the site since the conformation, size and chemical composition of the protein surface are all relevant for the interaction with a specific ligand. In our group, we designed an innovative bioinformatics method called SuMo in order to detect similar 3-dimensional (3D) sites in proteins (Jambon et al. Protein-Struct Funct Genet 52:137–145, 2003). This approach allowed the comparison of protein structures or substructures, and detected local spatial similarities: the main advantage of the method is its independence for both amino acid sequences and backbone structures. In contrast to already existing tools, the basis for this method is a representation of the protein structure by a set of stereo chemical groups that are defined independently from the notion of amino acid. An efficient heuristics for finding similarities has been developed which uses graphs of triangles of chemical groups to represent the protein structures. The SuMo (Surfing the Molecules) program allows the dynamic definition of chemical groups, the selection of sites in the proteins, and the management and screening of databases. The basic principle of SuMo has been used in several recent studies (Sperandio et al. J Cheml Inf Model 47:1097–1110, 2007) (Doppelt-Azeroual et al. Protein Sci 19:847–867, 2010). In order to give access to the SuMo tool, we proposed a web server (Jambon et al. Bioinformatics 21:3929–3930, 2005) reachable at http://sumo-pbil.ibcp.fr. This chapter will describe the main rationale we initially took for designing the first release of SuMo. In addition, we propose a completely new set of parameters best suitable for proteins and finally, we illustrate its power with several biological examples. Two of them dealing with serine proteases and lectins are given for a comparison purpose. The first two examples illustrate the capability of SuMo to deal with completely opposite modes of evolution i.e. convergence and divergence. A new biological application dealing with betalactame binding protein PBB molecules is also presented.