A5041591967- Computational Drug Discovery Methods
- Protein Structure and Dynamics
- Machine Learning in Materials Science
- Machine Learning in Bioinformatics
- Monoclonal and Polyclonal Antibodies Research
- RNA Research and Splicing
- Innovative Microfluidic and Catalytic Techniques Innovation
- Advanced Electron Microscopy Techniques and Applications
- Synthesis and Biological Evaluation
- Cell Image Analysis Techniques
- Biochemical and Structural Characterization
- Cancer-related gene regulation
- Genomics and Chromatin Dynamics
- Microbial Natural Products and Biosynthesis
- RNA and protein synthesis mechanisms
École Polytechnique Fédérale de Lausanne
2022-2024
SIB Swiss Institute of Bioinformatics
2023
Institut polytechnique de Grenoble
2020-2022
Université Grenoble Alpes
2020-2022
Centre National de la Recherche Scientifique
2020-2022
Laboratoire Jean Kuntzmann
2021-2022
Moscow Institute of Physics and Technology
2020-2021
Centre Inria de l'Université Grenoble Alpes
2020-2021
Ministry of Industry and Information Technology
2020
Abstract Fragment-based drug discovery has been an effective paradigm in early-stage development. An open challenge this area is designing linkers between disconnected molecular fragments of interest to obtain chemically relevant candidate molecules. In work, we propose DiffLinker, E(3)-equivariant three-dimensional conditional diffusion model for linker design. Given a set fragments, our places missing atoms and designs molecule incorporating all the initial fragments. Unlike previous...
Structure-based drug design (SBDD) aims to small-molecule ligands that bind with high affinity and specificity pre-determined protein targets. In this paper, we formulate SBDD as a 3D-conditional generation problem present DiffSBDD, an SE(3)-equivariant diffusion model generates novel conditioned on pockets. Comprehensive in silico experiments demonstrate the efficiency effectiveness of DiffSBDD generating diverse drug-like competitive docking scores. We further explore flexibility framework...
Structure-based drug design (SBDD) aims to small-molecule ligands that bind with high affinity and specificity pre-determined protein targets. Generative SBDD methods leverage structural data of drugs their targets propose new candidates. However, most existing focus exclusively on bottom-up de novo compounds or tackle other development challenges task-specific models. The latter requires curation suitable datasets, careful engineering the models retraining from scratch for each task. Here...
Fragment-based drug discovery has been an effective paradigm in early-stage development. An open challenge this area is designing linkers between disconnected molecular fragments of interest to obtain chemically-relevant candidate molecules. In work, we propose DiffLinker, E(3)-equivariant 3D-conditional diffusion model for linker design. Given a set fragments, our places missing atoms and designs molecule incorporating all the initial fragments. Unlike previous approaches that are only able...
Effective use of evolutionary information has recently led to tremendous progress in computational prediction three-dimensional (3D) structures proteins and their complexes. Despite the progress, accuracy predicted tends vary considerably from case case. Since utility models depends on accuracy, reliable estimates deviation between native are utmost importance.For first time, we present a deep convolutional neural network (CNN) constructed Voronoi tessellation 3D molecular structures....
Abstract Critical Assessment of Structure Prediction (CASP) is an organization aimed at advancing the state art in computing protein structure from sequence. In spring 2020, CASP launched a community project to compute structures most structurally challenging proteins coded for SARS‐CoV‐2 genome. Forty‐seven research groups submitted over 3000 three‐dimensional models and 700 sets accuracy estimates on 10 proteins. The resulting were released public. members also worked together provide...
Processing information on 3D objects requires methods stable to rigid-body transformations, in particular rotations, of the input data. In image processing tasks, convolutional neural networks achieve this property using rotation-equivariant operations. However, contrary images, graphs generally have irregular topology. This makes it challenging define a convolution operation these structures. work, we propose Spherical Graph Convolutional Network (S-GCN) that processes models proteins...
Motivation Effective use of evolutionary information has recently led to tremendous progress in computational prediction three-dimensional (3D) structures proteins and their complexes. Despite the progress, accuracy predicted tends vary considerably from case case. Since utility models depends on accuracy, reliable estimates deviation between native are utmost importance. Results For first time we present a deep convolutional neural network (CNN) constructed Voronoi tessellation 3D molecular...
A bstract Small molecules have been the preferred modality for drug development and therapeutic interventions. This molecular format presents a number of advantages, e.g. long half-lives cell permeability, making it possible to access wide range targets. However, finding small that engage “hard-to-drug” protein targets specifically potently remains an arduous process, requiring experimental screening extensive compound libraries identify candidate leads. The search continues with further...
Retrosynthesis planning is a fundamental challenge in chemistry which aims at designing reaction pathways from commercially available starting materials to target molecule. Each step multi-step retrosynthesis requires accurate prediction of possible precursor molecules given the molecule and confidence estimates guide heuristic search algorithms. We model single-step as distribution learning problem discrete state space. First, we introduce Markov Bridge Model, generative framework aimed...
In this work, we introduce 6D Convolutional Neural Network (6DCNN) designed to tackle the problem of detecting relative positions and orientations local patterns when processing three-dimensional volumetric data. 6DCNN also includes SE(3)-equivariant message-passing nonlinear activation operations constructed in Fourier space. Working space allows significantly reducing computational complexity our operations. We demonstrate properties convolution its efficiency recognition spatial patterns....