Raphael Bourgeas
Raphael Bourgeas has a total of 41 co-authors. Here are the top co-authors by number of shared publications:
- Philippe Roche
Centre de Recherche en Cancérologie de Marseille, Centre National de la Recherche Scientifique Unité Mixte de Recherche 7258, Institut National de la Santé et de la Recherche Médicale Unité 1068, Institut Paoli-Calmettes, and Aix-Marseille Université, 13009 Marseille, France; and
5 shared publications - Xavier Morelli
Centre de Recherche en Cancérologie de Marseille (CRCM), CNRS, INSERM, Aix-Marseille Univ, Institut Paoli-Calmettes, Marseille, France
4 shared publications - Fabrizio Pucci
Computational Biology and Bioinformatics, Université Libre de Bruxelles, 1050 Brussels, Belgium
3 shared publications - Marianne Rooman
Computational Biology and Bioinformatics, Université Libre de Bruxelles, Brussels, Belgium
3 shared publications - José Antonio Márquez
European Molecular Biology Laboratory, EMBL Grenoble, 71 avenue des Martyrs, 9018 38042 Grenoble, France
2 shared publications Marie Jeanne Basse
Laboratory of integrative Structural and Chemical Biology (iSCB), Centre de Recherche en Cancérologie de Marseille (CRCM); CNRS UMR 7258, INSERM U 1068, Institut Paoli-Calmettes; and Aix-Marseille Universités, Marseille 13009, France
2 shared publications
Latest Publications
- 31 March 2026
- Published by Agence Bibliographique de l'Enseignement Supérieur
2P2IDB : Une base de données dédiée à la druggabilité des interactions protéine-protéine.
Le nombre considérable d'interactions protéine-protéine (PPIs) existant au sein d'un organisme, ainsi que leur implication cruciale dans la vie cellulaire et dans de nombreuses pathologies, font des PPIs un immense réservoir de cibles potentielles pour la recherche de médicaments. Les PPIs sont aujourd'hui sur le devant de la scène grâce au développement de méthodologies innovantes et la validation récente de molécules chimiques modulant ces interactions dans des essais précliniques.L'étude des modulateurs d'interactions protéine-protéine (PPIM), a des implications tant dans la recherche fondamentale que thérapeutique. Les PPIMs peuvent aider à la compréhension des réseaux d'interactions. Elles permettront également de faire émerger de nouvelles familles d'agents thérapeutiques actifs dans diverses pathologies.Mon travail de thèse a principalement porté sur deux aspects de l'étude de l'inhibition des PPIs. D'une part, l'étude de l'implication des divers paramètres physicochimiques gouvernant une PPI dans sa capacité à être modulée (étude dite de la « druggabilité »), m'a amené à participer à la création d'une base de données structurale des interactions protéine-protéine : 2P2IDB (http://2p2idb.cnrs-mrs.fr/). D'autre part, j'ai contribué à l analyse de l'espace chimique des molécules présentes dans la base de données 2P2IDB. Nous avons défini la « Rule Of 4 » comme ligne de conduite pour caractériser ces molécules. Nous avons de plus utilisé le SVM afin de créer un protocole innovant (2P2IHUNTER) qui nous a permis de filtrer de grandes collections de composés afin de créer des chimiothèques dédiées aux PPIs.
- 31 March 2026
- Published by Agence Bibliographique de l'Enseignement Supérieur
2P2IDB : Une base de données dédiée à la druggabilité des interactions protéine-protéine.
Le nombre considérable d'interactions protéine-protéine (PPIs) existant au sein d'un organisme, ainsi que leur implication cruciale dans la vie cellulaire et dans de nombreuses pathologies, font des PPIs un immense réservoir de cibles potentielles pour la recherche de médicaments. Les PPIs sont aujourd'hui sur le devant de la scène grâce au développement de méthodologies innovantes et la validation récente de molécules chimiques modulant ces interactions dans des essais précliniques.L'étude des modulateurs d'interactions protéine-protéine (PPIM), a des implications tant dans la recherche fondamentale que thérapeutique. Les PPIMs peuvent aider à la compréhension des réseaux d'interactions. Elles permettront également de faire émerger de nouvelles familles d'agents thérapeutiques actifs dans diverses pathologies.Mon travail de thèse a principalement porté sur deux aspects de l'étude de l'inhibition des PPIs. D'une part, l'étude de l'implication des divers paramètres physicochimiques gouvernant une PPI dans sa capacité à être modulée (étude dite de la « druggabilité »), m'a amené à participer à la création d'une base de données structurale des interactions protéine-protéine : 2P2IDB (http://2p2idb.cnrs-mrs.fr/). D'autre part, j'ai contribué à l analyse de l'espace chimique des molécules présentes dans la base de données 2P2IDB. Nous avons défini la « Rule Of 4 » comme ligne de conduite pour caractériser ces molécules. Nous avons de plus utilisé le SVM afin de créer un protocole innovant (2P2IHUNTER) qui nous a permis de filtrer de grandes collections de composés afin de créer des chimiothèques dédiées aux PPIs.
- Open Access
- 19 March 2026
- Published by International Union of Crystallography (IUCr) in IUCrJ
- Vol. 13 (3) , 235-248
AXIS: a Lab-in-the-Loop machine learning approach for automated detection of macromolecular crystals
Macromolecular crystallography provides mechanistic understanding of biological processes and can be applied in drug design. Nowadays, the use of robotic systems for crystal growth and diffraction analysis is widespread and high-throughput protein-to-structure pipelines for ligand and fragment screening are revolutionizing the field. However, the identification of crystals is still largely carried out through manual inspection, sometimes involving tens of thousands of images, which represents a bottleneck in an otherwise highly automated process. Here we describe AXIS, an AI-based Crystal Identification System combining the DINOv2 computer vision model, state-of-the-art transfer learning and MARCO, the largest crystallization dataset available to date, for automated crystal detection. AXIS can operate with both visible and UV light images and integrates a Lab-in-the-Loop approach combining ML and expert inputs for iterative learning and specialization. AXIS enables automated annotation of large crystallization image datasets with performance and accuracy comparable to that of human experts, and the Lab-in-the-Loop approach introduced here enables efficient adaptation to local conditions, facilitating widespread application, which has been a major limitation to date. AXIS can help to correct human errors in image annotation and removes critical bottlenecks, particularly in the context of extensive crystallization screens or high-throughput applications like fragment and ligand screening, unlocking the potential for higher levels of automation that are key in both fundamental and translational research.
- 19 March 2026
- Published by International Union of Crystallography (IUCr) in IUCrJ
- Vol. 13 (3) , 235-248
Open AccessAXIS: a Lab-in-the-Loop machine learning approach for automated detection of macromolecular crystals
Macromolecular crystallography provides mechanistic understanding of biological processes and can be applied in drug design. Nowadays, the use of robotic systems for crystal growth and diffraction analysis is widespread and high-throughput protein-to-structure pipelines for ligand and fragment screening are revolutionizing the field. However, the identification of crystals is still largely carried out through manual inspection, sometimes involving tens of thousands of images, which represents a bottleneck in an otherwise highly automated process. Here we describe AXIS, an AI-based Crystal Identification System combining the DINOv2 computer vision model, state-of-the-art transfer learning and MARCO, the largest crystallization dataset available to date, for automated crystal detection. AXIS can operate with both visible and UV light images and integrates a Lab-in-the-Loop approach combining ML and expert inputs for iterative learning and specialization. AXIS enables automated annotation of large crystallization image datasets with performance and accuracy comparable to that of human experts, and the Lab-in-the-Loop approach introduced here enables efficient adaptation to local conditions, facilitating widespread application, which has been a major limitation to date. AXIS can help to correct human errors in image annotation and removes critical bottlenecks, particularly in the context of extensive crystallization screens or high-throughput applications like fragment and ligand screening, unlocking the potential for higher levels of automation that are key in both fundamental and translational research.
- Open Access
It’s Not Always Crystal Clear: In-Pocket Analysis for Understanding Binding Site Microenvironment Properties
Structural biology enabled key advances in biochemistry, drug discovery, and AI-driven three-dimensional structure modelling. Despite its importance, reliable and general computational tools for chemical validation of structural models remain limited, especially when small-molecule ligands are involved. This challenge is exacerbated when experimental resolution is insufficient to define interaction chemistry, leaving critical interpretations ambiguous. We introduce In-Pocket, a quantum mechanical optimisation method biased by experimental data that evaluates the chemical nature of protein-ligand binding interfaces. In-Pocket improves model quality, corrects chemically inconsistent structures, and identifies the energetically most favourable ligand-conformers within biomacromolecular complexes. The approach is generally applicable to a wide range of systems, including those containing transition metals. Benchmarking demonstrates that In-Pocket enhances the reliability of legacy and newly solved structures. With automated, user-friendly routine chemical validation, In-Pocket offers unique opportunities to improve the integrity of structural databases and impact diverse domains: from molecular biology to rational drug design and application of generative AI.
Open AccessIt’s Not Always Crystal Clear: In-Pocket Analysis for Understanding Binding Site Microenvironment Properties
Structural biology enabled key advances in biochemistry, drug discovery, and AI-driven three-dimensional structure modelling. Despite its importance, reliable and general computational tools for chemical validation of structural models remain limited, especially when small-molecule ligands are involved. This challenge is exacerbated when experimental resolution is insufficient to define interaction chemistry, leaving critical interpretations ambiguous. We introduce In-Pocket, a quantum mechanical optimisation method biased by experimental data that evaluates the chemical nature of protein-ligand binding interfaces. In-Pocket improves model quality, corrects chemically inconsistent structures, and identifies the energetically most favourable ligand-conformers within biomacromolecular complexes. The approach is generally applicable to a wide range of systems, including those containing transition metals. Benchmarking demonstrates that In-Pocket enhances the reliability of legacy and newly solved structures. With automated, user-friendly routine chemical validation, In-Pocket offers unique opportunities to improve the integrity of structural databases and impact diverse domains: from molecular biology to rational drug design and application of generative AI.
- 1 June 2016
- Published by AIP Publishing in Journal of Physical and Chemical Reference Data
- Vol. 45 (2) , 023104
High-quality Thermodynamic Data on the Stability Changes of Proteins Upon Single-site Mutations
We have set up and manually curated a dataset containing experimental information on the impact of amino acid substitutions in a protein on its thermal stability. It consists of a repository of experimentally measured melting temperatures (Tm) and their changes upon point mutations (ΔTm) for proteins having a well-resolved x-ray structure. This high-quality dataset is designed for being used for the training or benchmarking of in silico thermal stability prediction methods. It also reports other experimentally measured thermodynamic quantities when available, i.e., the folding enthalpy (ΔH) and heat capacity (ΔCP) of the wild type proteins and their changes upon mutations (ΔΔH and ΔΔCP), as well as the change in folding free energy (ΔΔG) at a reference temperature. These data are analyzed in view of improving our insights into the correlation between thermal and thermodynamic stabilities, the asymmetry between the number of stabilizing and destabilizing mutations, and the difference in stabilization potential of thermostable versus mesostable proteins.
- 1 June 2016
- Published by AIP Publishing in Journal of Physical and Chemical Reference Data
- Vol. 45 (2) , 023104
High-quality Thermodynamic Data on the Stability Changes of Proteins Upon Single-site Mutations
We have set up and manually curated a dataset containing experimental information on the impact of amino acid substitutions in a protein on its thermal stability. It consists of a repository of experimentally measured melting temperatures (Tm) and their changes upon point mutations (ΔTm) for proteins having a well-resolved x-ray structure. This high-quality dataset is designed for being used for the training or benchmarking of in silico thermal stability prediction methods. It also reports other experimentally measured thermodynamic quantities when available, i.e., the folding enthalpy (ΔH) and heat capacity (ΔCP) of the wild type proteins and their changes upon mutations (ΔΔH and ΔΔCP), as well as the change in folding free energy (ΔΔG) at a reference temperature. These data are analyzed in view of improving our insights into the correlation between thermal and thermodynamic stabilities, the asymmetry between the number of stabilizing and destabilizing mutations, and the difference in stabilization potential of thermostable versus mesostable proteins.
- Open Access
- 18 March 2016
- Published by Springer Nature in Scientific Reports
- Vol. 6 (1) , 23257
Predicting protein thermal stability changes upon point mutations using statistical potentials: Introducing HoTMuSiC
The accurate prediction of the impact of an amino acid substitution on the thermal stability of a protein is a central issue in protein science, and is of key relevance for the rational optimization of various bioprocesses that use enzymes in unusual conditions. Here we present one of the first computational tools to predict the change in melting temperature ΔTm upon point mutations, given the protein structure and, when available, the melting temperature Tm of the wild-type protein. The key ingredients of our model structure are standard and temperature-dependent statistical potentials, which are combined with the help of an artificial neural network. The model structure was chosen on the basis of a detailed thermodynamic analysis of the system. The parameters of the model were identified on a set of more than 1,600 mutations with experimentally measured ΔTm. The performance of our method was tested using a strict 5-fold cross-validation procedure, and was found to be significantly superior to that of competing methods. We obtained a root mean square deviation between predicted and experimental ΔTm values of 4.2 °C that reduces to 2.9 °C when ten percent outliers are removed. A webserver-based tool is freely available for non-commercial use at soft.dezyme.com.
- 18 March 2016
- Published by Springer Nature in Scientific Reports
- Vol. 6 (1) , 23257
Open AccessPredicting protein thermal stability changes upon point mutations using statistical potentials: Introducing HoTMuSiC
The accurate prediction of the impact of an amino acid substitution on the thermal stability of a protein is a central issue in protein science, and is of key relevance for the rational optimization of various bioprocesses that use enzymes in unusual conditions. Here we present one of the first computational tools to predict the change in melting temperature ΔTm upon point mutations, given the protein structure and, when available, the melting temperature Tm of the wild-type protein. The key ingredients of our model structure are standard and temperature-dependent statistical potentials, which are combined with the help of an artificial neural network. The model structure was chosen on the basis of a detailed thermodynamic analysis of the system. The parameters of the model were identified on a set of more than 1,600 mutations with experimentally measured ΔTm. The performance of our method was tested using a strict 5-fold cross-validation procedure, and was found to be significantly superior to that of competing methods. We obtained a root mean square deviation between predicted and experimental ΔTm values of 4.2 °C that reduces to 2.9 °C when ten percent outliers are removed. A webserver-based tool is freely available for non-commercial use at soft.dezyme.com.
- Open Access
High-quality thermodynamic data on the stability changes of proteins upon single-site mutations
We have set up and manually curated a dataset containing experimental information on the impact of amino acid substitutions in a protein on its thermal stability. It consists of a repository of experimentally measured melting temperatures ( T m ) and their changes upon point mutations (Δ T m ) for proteins having a well-resolved X-ray structure. This high-quality dataset is designed for being used for the training or benchmarking of in silico thermal stability prediction methods. It also reports other experimentally measured thermodynamic quantities when available, i.e . the folding enthalpy (Δ H ) and heat capacity (Δ C P ) of the wild type proteins and their changes upon mutations (ΔΔ H and ΔΔ C P ), as well as the change in folding free energy (ΔΔ G ) at a reference temperature. These data are analyzed in view of improving our insights into the correlation between thermal and thermodynamic stabilities, the asymmetry between the number of stabilizing and destabilizing mutations, and the difference in stabilization potential of thermostable versus mesostable proteins.
Open AccessHigh-quality thermodynamic data on the stability changes of proteins upon single-site mutations
We have set up and manually curated a dataset containing experimental information on the impact of amino acid substitutions in a protein on its thermal stability. It consists of a repository of experimentally measured melting temperatures ( T m ) and their changes upon point mutations (Δ T m ) for proteins having a well-resolved X-ray structure. This high-quality dataset is designed for being used for the training or benchmarking of in silico thermal stability prediction methods. It also reports other experimentally measured thermodynamic quantities when available, i.e . the folding enthalpy (Δ H ) and heat capacity (Δ C P ) of the wild type proteins and their changes upon mutations (ΔΔ H and ΔΔ C P ), as well as the change in folding free energy (ΔΔ G ) at a reference temperature. These data are analyzed in view of improving our insights into the correlation between thermal and thermodynamic stabilities, the asymmetry between the number of stabilizing and destabilizing mutations, and the difference in stabilization potential of thermostable versus mesostable proteins.
- Open Access
- 6 January 2014
- Published by The Royal Society in Journal of The Royal Society Interface
- Vol. 11 (90) , 20130860
2P2I HUNTER : a tool for filtering orthosteric protein–protein interaction modulators via a dedicated support vector machine
Over the last 10 years, protein–protein interactions (PPIs) have shown increasing potential as new therapeutic targets. As a consequence, PPIs are today the most screened target class in high-throughput screening (HTS). The development of broad chemical libraries dedicated to these particular targets is essential; however, the chemical space associated with this ‘high-hanging fruit’ is still under debate. Here, we analyse the properties of 40 non-redundant small molecules present in the 2P2I database ( http://2p2idb.cnrs-mrs.fr/ ) to define a general profile of orthosteric inhibitors and propose an original protocol to filter general screening libraries using a support vector machine (SVM) with 11 standard D ragon molecular descriptors. The filtering protocol has been validated using external datasets from PubChem BioAssay and results from in-house screening campaigns . This external blind validation demonstrated the ability of the SVM model to reduce the size of the filtered chemical library by eliminating up to 96% of the compounds as well as enhancing the proportion of active compounds by up to a factor of 8. We believe that the resulting chemical space identified in this paper will provide the scientific community with a concrete support to search for PPI inhibitors during HTS campaigns.
- 6 January 2014
- Published by The Royal Society in Journal of The Royal Society Interface
- Vol. 11 (90) , 20130860
Open Access2P2I HUNTER : a tool for filtering orthosteric protein–protein interaction modulators via a dedicated support vector machine
Over the last 10 years, protein–protein interactions (PPIs) have shown increasing potential as new therapeutic targets. As a consequence, PPIs are today the most screened target class in high-throughput screening (HTS). The development of broad chemical libraries dedicated to these particular targets is essential; however, the chemical space associated with this ‘high-hanging fruit’ is still under debate. Here, we analyse the properties of 40 non-redundant small molecules present in the 2P2I database ( http://2p2idb.cnrs-mrs.fr/ ) to define a general profile of orthosteric inhibitors and propose an original protocol to filter general screening libraries using a support vector machine (SVM) with 11 standard D ragon molecular descriptors. The filtering protocol has been validated using external datasets from PubChem BioAssay and results from in-house screening campaigns . This external blind validation demonstrated the ability of the SVM model to reduce the size of the filtered chemical library by eliminating up to 96% of the compounds as well as enhancing the proportion of active compounds by up to a factor of 8. We believe that the resulting chemical space identified in this paper will provide the scientific community with a concrete support to search for PPI inhibitors during HTS campaigns.
- Open Access
- 30 November 2012
- Published by Oxford University Press (OUP) in Nucleic Acids Research
- Vol. 41 (D1) , D824-D827
2P2Idb: a structural database dedicated to orthosteric modulation of protein–protein interactions
Protein–protein interactions are considered as one of the next generation of therapeutic targets. Specific tools thus need to be developed to tackle this challenging chemical space. In an effort to derive some common principles from recent successes, we have built 2P2Idb (freely accessible at http://2p2idb.cnrs-mrs.fr), a hand-curated structural database dedicated to protein–protein interactions with known orthosteric modulators. It includes all interactions for which both the protein–protein and protein–ligand complexes have been structurally characterized. A web server provides links to related sites of interest, binding affinity data, pre-calculated structural information about protein–protein interfaces and 3D interactive views through java applets. Comparison of interfaces in 2P2Idb to those of representative datasets of heterodimeric complexes has led to the identification of geometrical parameters and residue properties to assess the druggability of protein–protein complexes. A tool is proposed to calculate a series of biophysical and geometrical parameters that characterize protein–protein interfaces. A large range of descriptors are computed including, buried accessible surface area, gap volume, non-bonded contacts, hydrogen-bonds, atom and residue composition, number of segments and secondary structure contribution. All together the 2P2I database represents a structural source of information for scientists from academic institutions or pharmaceutical industries.
- 30 November 2012
- Published by Oxford University Press (OUP) in Nucleic Acids Research
- Vol. 41 (D1) , D824-D827
Open Access2P2Idb: a structural database dedicated to orthosteric modulation of protein–protein interactions
Protein–protein interactions are considered as one of the next generation of therapeutic targets. Specific tools thus need to be developed to tackle this challenging chemical space. In an effort to derive some common principles from recent successes, we have built 2P2Idb (freely accessible at http://2p2idb.cnrs-mrs.fr), a hand-curated structural database dedicated to protein–protein interactions with known orthosteric modulators. It includes all interactions for which both the protein–protein and protein–ligand complexes have been structurally characterized. A web server provides links to related sites of interest, binding affinity data, pre-calculated structural information about protein–protein interfaces and 3D interactive views through java applets. Comparison of interfaces in 2P2Idb to those of representative datasets of heterodimeric complexes has led to the identification of geometrical parameters and residue properties to assess the druggability of protein–protein complexes. A tool is proposed to calculate a series of biophysical and geometrical parameters that characterize protein–protein interfaces. A large range of descriptors are computed including, buried accessible surface area, gap volume, non-bonded contacts, hydrogen-bonds, atom and residue composition, number of segments and secondary structure contribution. All together the 2P2I database represents a structural source of information for scientists from academic institutions or pharmaceutical industries.
- 1 August 2011
- Published by Elsevier in Current Opinion in Chemical Biology
- Vol. 15 (4) , 475-481
Chemical and structural lessons from recent successes in protein–protein interaction inhibition (2P2I)
- 1 August 2011
- Published by Elsevier in Current Opinion in Chemical Biology
- Vol. 15 (4) , 475-481
Chemical and structural lessons from recent successes in protein–protein interaction inhibition (2P2I)
- Open Access
- 9 March 2010
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 5 (3) , e9598
Atomic Analysis of Protein-Protein Interfaces with Known Inhibitors: The 2P2I Database
In the last decade, the inhibition of protein-protein interactions (PPIs) has emerged from both academic and private research as a new way to modulate the activity of proteins. Inhibitors of these original interactions are certainly the next generation of highly innovative drugs that will reach the market in the next decade. However, in silico design of such compounds still remains challenging. Here we describe this particular PPI chemical space through the presentation of 2P2IDB, a hand-curated database dedicated to the structure of PPIs with known inhibitors. We have analyzed protein/protein and protein/inhibitor interfaces in terms of geometrical parameters, atom and residue properties, buried accessible surface area and other biophysical parameters. The interfaces found in 2P2IDB were then compared to those of representative datasets of heterodimeric complexes. We propose a new classification of PPIs with known inhibitors into two classes depending on the number of segments present at the interface and corresponding to either a single secondary structure element or to a more globular interacting domain. 2P2IDB complexes share global shape properties with standard transient heterodimer complexes, but their accessible surface areas are significantly smaller. No major conformational changes are seen between the different states of the proteins. The interfaces are more hydrophobic than general PPI's interfaces, with less charged residues and more non-polar atoms. Finally, fifty percent of the complexes in the 2P2IDB dataset possess more hydrogen bonds than typical protein-protein complexes. Potential areas of study for the future are proposed, which include a new classification system consisting of specific families and the identification of PPI targets with high druggability potential based on key descriptors of the interaction. 2P2I database stores structural information about PPIs with known inhibitors and provides a useful tool for biologists to assess the potential druggability of their interfaces. The database can be accessed at http://2p2idb.cnrs-mrs.fr.
- 9 March 2010
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 5 (3) , e9598
Open AccessAtomic Analysis of Protein-Protein Interfaces with Known Inhibitors: The 2P2I Database
In the last decade, the inhibition of protein-protein interactions (PPIs) has emerged from both academic and private research as a new way to modulate the activity of proteins. Inhibitors of these original interactions are certainly the next generation of highly innovative drugs that will reach the market in the next decade. However, in silico design of such compounds still remains challenging. Here we describe this particular PPI chemical space through the presentation of 2P2IDB, a hand-curated database dedicated to the structure of PPIs with known inhibitors. We have analyzed protein/protein and protein/inhibitor interfaces in terms of geometrical parameters, atom and residue properties, buried accessible surface area and other biophysical parameters. The interfaces found in 2P2IDB were then compared to those of representative datasets of heterodimeric complexes. We propose a new classification of PPIs with known inhibitors into two classes depending on the number of segments present at the interface and corresponding to either a single secondary structure element or to a more globular interacting domain. 2P2IDB complexes share global shape properties with standard transient heterodimer complexes, but their accessible surface areas are significantly smaller. No major conformational changes are seen between the different states of the proteins. The interfaces are more hydrophobic than general PPI's interfaces, with less charged residues and more non-polar atoms. Finally, fifty percent of the complexes in the 2P2IDB dataset possess more hydrogen bonds than typical protein-protein complexes. Potential areas of study for the future are proposed, which include a new classification system consisting of specific families and the identification of PPI targets with high druggability potential based on key descriptors of the interaction. 2P2I database stores structural information about PPIs with known inhibitors and provides a useful tool for biologists to assess the potential druggability of their interfaces. The database can be accessed at http://2p2idb.cnrs-mrs.fr.