This Week In Cheminformatics: Issue #014
Guide to Mixed-Solvent MD and Solvent-Site-Biased Docking, StereoMolGraph and a long list of papers
Highlights
Best Practices in Mixed-Solvent Molecular Dynamics and Solvent-Site-Biased Docking
While mixed-solvent molecular dynamics and solvent-site-biased docking are excellent tools for mapping protein interaction hot spots, their practical implementations have historically been highly heterogeneous and ad hoc across the literature. This Perspective by Prieto et al. provides a much needed unified methodological framework, establishing concrete best practices. The paper advocates for water and ethanol as foundational probes, defines strict convergence criteria using probe finding probabilities, a 1-1.5 Å spatial cutoff for clustering, and details how to translate these thermodynamically grounded sites into discrete docking biases using inverted Gaussian potentials. Good read.
StereoMolGraph: Stereochemistry-Aware Molecular and Reaction Graphs
Handling complex stereochemistry is non trivial and anyone who has tried to build their own SMILES parser knows this very well. In this paper, Papusha and Leonhard introduce StereoMolGraph, an open-source Python library that bypasses the canonization process entirely. Instead, it encodes relative spatial arrangements using permutation-invariant local stereodescriptors governed by rotational symmetry groups. Structural equivalence, including the identification of exact enantiomers and diastereomers, is then evaluated pairwise via an extended VF2++ isomorphism algorithm that maps these stereodescriptors. It natively handles non-tetrahedral geometries like octahedral metal complexes and uses condensed reaction graphs to capture fleeting stereochemistry in transition states. Built with RDKit interoperability in mind, it provides a transparent, mathematical alternative for stereochemically sensitive workflows where standard detection algorithms might frequently fail.
Long List
Cheminformatics
Reaction optimization through mechanistic insight and predictive modelling
High-performance training and inference for deep equivariant interatomic potentials
FePTP: A text-mined dataset of transformation pathways among iron-containing phases
Exploring Secondary Structure Predictions for RNA-Targeted Drug Discovery: Power and Challenges
Why Is the Peak Group Analysis So Effective for IR Spectra Analysis?
TabPFN Opens New Avenues for Small-Data Tabular Learning in Drug Discovery
CHARMM-GUI Ligand Docker for Molecular Docking with Various Docking Programs
Pharmacokinetic Prediction of Repurposed Drugs for PDAC Using Artificial Intelligence
MXtalTools: A Toolkit for Machine Learning on Molecular Crystals
ProtCross: Bridging the PDB-AlphaFold Gap for Binding Site Prediction with Protein Point Clouds
Physics-Guided Machine Learning for Ionic-Liquid Volumetric Properties
AssayMatch: Learning To Select Data for Molecular Activity Models
Data-Driven Design Guidelines for TADF Emitters from a High-Throughput Screening of 747 Molecules
Reproducible Adaptive MCMC via Sharing a Pretrained Generator Matrix across Runs and Structures
A DNN Biophysics Model with Topological and Electrostatic Features
AutoPocket2CREST: Automating Binding Pocket Extraction for the CREST Conformer Generation Pipeline
MedChem
Other
Palate Cleanser
hehe,
Manas
