Project: A Systems Biology approach: Uniting protein knowledge to integrated structural analysis

Protein structure modeling and analysis is a powerful approach for understanding molecular interactions and protein functions. It has become invaluable in the pharmaceutical industry for rational drug design, from the lead discovery to the structural optimization of drug candidates. Favored by world-wide efforts in Structural Genomic projects, experimentally determined protein structures are rapidly increasing and already cover 21% of human proteins, whereas only 10 years ago, the cover was only of a few percents. Although some studies already estimate that as much as 66% of human proteins could be modeled by homology using structural templates (Pieper U et al, 2009), we estimate that a fair 35 to 40% of human protein structures can already be accurately modeled. This new availability of protein structures with an ever-growing cover of the human proteome is only recently giving rise to new integrated structural approaches (Higurashi M et al, 2009; Albou LP et al, 2010): Structural Biology is meeting Systems Biology by providing invaluable geometrical and physico-chemical information to understand and predict molecular interactions and recognition mechanisms._x000D__x000D_In this context, BIONEXT has created an entirely new approach based on contiguous 3D regions to perform the virtual screening (comparison) of molecular structures, including proteins, peptides, ligands and RNA, as well as the screening of biological assemblies? interfaces. The company has demonstrated its uses to detect novel anti-targets of drugs, completely unrelated in terms of sequence and structure to already known targets (International patents EP2010/060821, EP2010/060822, Albou LP et al, 2008-2010). BIONEXT is now actively engaged in the early assessment of the functions, side-effects, toxicity and in vivo efficacy of molecules at every stage of their development. Those phenomena represent approximately 50% of candidate drugs failures and are responsible for a success rate of 1 for 8000 candidates with an average development time of 10-15 years and costs of nearly 1 billion euros (Business Insights, 2009). Recent news also relate numerous cases of drugs either stopped or in surveillance due to these phenomena._x000D__x000D_The power of the BVS technology to answer these questions can however be considerably improved by the integration of reliable, informative biological data. The present project thus proposes to build synergies between BIONEXT, the Swiss Institute of Bioinformatics (SIB) and GeneBio, in order to bridge the gap between integrated structural analysis and protein knowledge. We will make use of the high-quality manual protein annotations from the neXtProt knowledge platform (http://beta.neXtProt.org/), developed by the SIB and GeneBio, to enrich our integrated analysis of structural data with extra biological context. Taking advantage of both structural data and protein knowledge, the modeling of cell and organelle environments will be undertaken to enhance the analysis and predictions of molecular interactions, recognition specificities, side-effects and in vivo efficacy inside these specific biological environments. To this end, subcellular localization information will be used to create organelle-specific databases of 3D contiguous regions and will provide important evidence to assess whether two molecules having the chemical ability to interact are indeed biological Ps. For instance, the mitochondrial environment has taken an increasing importance due to drug-induced toxicities (J.A.Daykens et al, 2007). Additionally, data related to expression in different tissues or cell types will also be integrated and will enrich the current databases of 3D contiguous regions by providing the users with cell-specific databases, reflecting the available molecular components and interactions for each given cell environment. _x000D__x000D_Furthermore, the BVS strategy of screening will also be used to predict protein Post-Translation Modifications (PTM) that play an essential role in the dynamic regulations of protein interactions. The collaboration between BIONEXT, the SIB and GeneBio is essential to develop those high-quality predictive approaches and is expected to improve the understanding and prediction of in vivo regulatory mechanisms. _x000D_The products developed in this project will benefit GeneBio, the SIB and BIONEXT by adding predictive approaches towards the in vivo accurate description of molecular interactions and their related effects in cells and organelles. Both BIONEXT and GeneBio will be able to commercialize those integrated solutions in their respective BVS and neXtProt platforms. Additionally, while the neXtProt users will be able to access reliable results from highly complex protein structure analysis, it will give BIONEXT the opportunity to present BVS to a large potential clientele._x000D_

Acronym	BioNeXtProt (Reference Number: 6715)
Duration	01/09/2011 - 01/02/2014
Project Topic	BioNeXtProt is a collaborative project between BIONEXT the SIB and GeneBio, uniting experts in protein knowledge and structural analysis to develop and commercialize high-quality predictive approaches for structure-based drug design and biological systems explorations and understanding.
Project Results (after finalisation)	The BioNeXtProt was a collaborative project designed to unite experts in Integrated Structure Analysis and Drug Design (BIONEXT) and experts in Protein Knowledge and System Biology (SIB and GeneBio). The CO focus of this project was to deliver analytical and predictive software to better predict and characterize the interactions between molecules inside complex cells or organelles. More precisely, the 4 CO goals of the project which are successful were :_x000D_ _x000D_$ the modeling of 300 cell types and organelles thanks to Structural, Proteomic and Antibody (HPA) data, to determine the presence/absence of more than 60% of all human proteins_x000D__x000D_$ the prediction at cellular/organelle scale of protein-protein and drug-target interactions, thanks to the Biological Virtual Simulation (BVS) platform of BIONEXT and the modeling of cells & organelles_x000D__x000D_$ the prediction at the proteome scale of Post-Translational Modifications (PTM) that can dynamically regulate the protein-protein and drug-target interactions. In particular in the case of phosphorylations which are well known to regulate molecular interactions and signaling pathways_x000D__x000D_$ developing & integrating a free software version of our Molecular Structure Visualizer (L-MSVM) inside the database neXtProt. L-MSVM benefits to SIB & GeneBio by providing a state-of-the-art visualization software easy to interface, while BIONEXT benefits from the already existing users of neXtProt to provide them with easy-to-use commercial applications to predict protein-protein and drug-target interactions through BVS_x000D__x000D_The two technical challenges relied 1) on the integration of heterogeneous big biological data which was tackled thanks to the expertise of the 3 Ps, and 2) on the ability for BVS users to easily launch online complex analysis from the neXtProt database._x000D__x000D__x000D_
Network	Eurostars
Call	Eurostars Cut-Off 6

Project partner

Number	Name	Role	Country
3	BIONEXT	Coordinator	France
3	Geneva Bioinformatics (GeneBio) SA	Partner	Switzerland
3	Swiss Institute of Bioinformatics	Partner	Switzerland