Mehdi D. Davari's Projects
Auto in silico Consensus Inverse Docking(ACID) is a web server mainly for drug repurposing based on the consensus inverse docking method, which is designed to evaluate the binding affinities between each protein and the given small molecule. ACID is open to the public and shows great potential.
Quantitative prediction of mutant function across an enzyme family
ChemicalTagger is a tool for semantic text-mining in chemistry.
Democratizing Deep-Learning for Drug Discovery, Quantum Chemistry, Materials Science and Biology
An experimental package for deep learning for molecular docking
Example code from the book "Deep Learning for the Life Sciences"
A generative latent variable model for biological sequence families.
Protocols and tools to run (automated) atomistic simulations of enzyme-ligand systems
Code for the paper "Enzyme Promiscuity Prediction using hierarchy-informed multi-label classification"
Evolutionary couplings from protein and RNA sequence alignments
Mutation effects predicted from sequence co-variation
Visually explore covariation in protein families
Reimplementation of the UniRep protein featurization model.
Code and data to reproduce analyses in Biswas et al. (2020) "Low-N protein engineering with data-efficient deep learning".
A complete, open-source, end-to-end re-implementation of the Church Lab's low-N eUniRep in silico protein engineering pipeline presented in Biswas et al.
Listing of papers about machine learning for proteins.
This introduces basics of machine learning packages
Generative Models for Graph-Based Protein Design
Inference of couplings in proteins and RNAs from sequence variation
A hybrid approach to protein structure (dihedral / torsional angles) prediction using techniques from Mohammed AlQuraishi's work on End-to-end differentiable learning of protein structure and Gao et al. on RaptorX.
Collected scripts for Pymol
PyPEF – Pythonic Protein Engineering Framework
Files and info for a practical session in the MSc Biophysics course run by Biochemistry
Reinforcement Learning based bioretrosynthesis tool
Tasks Assessing Protein Embeddings (TAPE), a set of five biologically relevant semi-supervised learning tasks spread across different domains of protein biology.
TAXyl (Thermal activity prediction for xylanase)
UniRep model, usage, and examples.