This Week In Cheminformatics: Issue #001

Holiday highlights you might have skipped. Plus, why I’m starting this.

Yet Another Newsletter...Why ??

Science Twitter has fragmented (you know why), and our feeds on LinkedIn and BlueSky have become unreliable for discovery. Finding high-quality research shouldn’t require navigating social media or sifting through humblebrags. This newsletter is simply an effort to keep track of relevant work without the noise of algorithmic feeds.

This Week in Cheminformatics is (supposed to be a) weekly curation of publications, preprints, code, and datasets. The scope covers cheminformatics, molecular modeling, and machine learning for drug discovery. The focus here is utility over engagement, so do with these links what your heart desires. Every Monday, I will share the previous week’s technical updates alongside any specific items that catch my eye.

Links

• Generative inverse design of RNA structure and function with gRNAde

• Chemprop v2: An Efficient, Modular Machine Learning Package for Chemical Property Prediction

• MambaTransDTA: A Hybrid Mamba-Transformer Architecture for Accurate Drug-Target Binding Affinity Prediction

• mzLearn as a data-driven LC/MS signal detection algorithm that enables pre-trained generative models for untargeted metabolomics

• ChemFM as a scaling law guided foundation model pre-trained on informative chemicals

• Multimodal out-of-distribution individual uncertainty quantification enhances binding affinity prediction for polypharmacology

• SynGFN: learning across chemical space with generative flow-based molecular discovery

• Site-Selective Protein Modification via Peptide-Directed Proximity Catalysis

• Constructing and explaining machine learning models for the exploration and design of boron-based Lewis acids

• Do Llamas understand the periodic table?

• GoFlow: efficient transition state geometry prediction with flow matching and E(3)-equivariant neural networks

• Generalized DeepONets for viscosity prediction using learned entropy scaling references

• Predicting PROTAC-mediated ternary complexes with AlphaFold3 and Boltz-1

• MolEncoder: towards optimal masked language modeling for molecules

• Leveraging large language models for enzymatic reaction prediction and characterization

• Database mining of ZINC15 natural compounds reveals potential thyroid receptor β agonists for NAFLD management: an in silico study

• Navigating materials design spaces with efficient Bayesian optimization: a case study in functionalized nanoporous materials

• Advancing mutagenicity predictions in drug discovery with an explainable few-shot deep learning framework

• Multi-modal contrastive learning for chemical structure elucidation with VibraCLIP

• Application-specific machine-learned interatomic potentials: exploring the trade-off between DFT convergence, MLIP expressivity, and computational cost

• Retrosynformer: planning multi-step chemical synthesis routes via a decision transformer

• Mol2Raman: a graph neural network model for predicting Raman spectra from SMILES representations

• SLAB: simultaneous labeling and binding affinity prediction for protein–ligand structures

• Hierarchical Attention Graph Learning with LLMs Enhancement for Molecular Solubility Prediction

• Multi-agentic AI framework for end-to-end atomistic simulations

• One step retrosynthesis of drugs from commercially available chemical building blocks and conceivable coupling reactions

• Explainable Active Learning Framework for Ligand Binding Affinity Prediction

• Computer vision for high-throughput materials synthesis: a tutorial for experimentalists

• Kinetic predictions for S_N2 reactions using the BERT architecture: Comparison and interpretation

• An exploration of dataset bias in single-step retrosynthesis prediction

• HGAlign: Biologically preserving batch correction and classification for metabolomics via heterogeneous graph alignment

• IFNg_DeepKG: A Novel Model for Identifying Interferon-Gamma-Inducing Epitopes Using Knowledge Graph RAG in Biomedical Applications

• One-Class Genetic Algorithm for Authentication Analysis of Spectrochemical Data

• Fractional Kinetic Modelling of the Adsorption and Desorption Processes From Experimental SPR Curves

• MetaSense: Enhancing Drug Metabolite Prediction with Structure-Aware Attention and Adaptive Diversity Beam Search

• Cross-Geometry Pretraining for Equivariant Neural Networks: Improving Molecular Property Prediction Under Computational Constraints

• Peptide-Tools–Web Server for Calculating Physicochemical Properties of Peptides

• Leveraging Consensus Docking Approaches for Human Mitochondrial Complexes I and III

• AlphaFold-RandomWalk and AlphaFold-Ensemble: Sampling Alternative Protein Conformations with Perturbed Versions of AlphaFold

• Path-Based Graph Neural Network for Drug Synergy Prediction and Interpretation

• Analytical Chemistry in Daily Life – ChemEye: A Smartphone Based Mobile App for Colorimetric Analysis

• Reliable Redox-Potential Simulations of Proteins

• Deciphering DNA’s Sequence-Dependent Structure and Deformability with Normalizing Flows

• Gaussian process emulation for exploring complex infectious disease models

• Cell-DINO: Self-supervised image-based embeddings for cell fluorescent microscopy

• A data-driven biology-based network model reproduces C. elegans premotor neural dynamics

• ViVo: A Temporal Modeling Framework That Boosts Statistical Power and Minimizes Animal Usage

• CalVSP: a program for analyzing the molecular surface areas, volumes, and polar surface areas

• Needle-in-a-haystack approach: rapid screening of PDE1C inhibitors through the combination of machine learning, molecular docking, molecular dynamics simulations and experimental validation

• ChemTSv3: Generalizing Molecular Design via Flexible Search Space Control

• StereoMolGraph: Stereochemistry-Aware Molecular and Reaction Graphs

• A Principal Component Analysis-Integrated Machine Learning Approach for Predicting Gas-Phase VUV/UV Absorption Spectra of Molecular Compounds

• EDWARD: E(3)-Equivariant Dual-Way Attentive Reduction for Peptide-to-Small-Molecule Design

• In silico identification of a prognostic gene signature and virtual screening for hepatocellular carcinoma

• scREPA: Predicting single-cell perturbation responses with cycle-consistent representation alignments

• PerturbSynX: Deep learning framework for predicting drug combination synergy scores using drug induced gene perturbation data

• MethylMSI: Prediction of microsatellite instability based on DNA methylation profile and SVM model

• RxnNet: An AI Framework for Reaction Mechanism Discovery - A Case Study of Carbocations

• Guiding Ligand Selection in Copper-Catalyzed Cross Couplings via Principal Component Analysis

• Graph-Based Internal Coordinate Analysis for Transition State Characterisation

• DeepMLP: A Proteomics-Driven Deep Learning Framework for Identifying Mis-Localized Proteins across Pan-Cancer

• AttentionScore: A Target-Specific, Bias-Aware Scoring Function for Structure-Based Virtual Screening: A Case Study on METTL3

• HetGAT-LMI: Interpretable Heterogeneous Graph Attention Method for Predicting lncRNA–miRNA Interactions

• Cluster-Graph Fingerprinting: A Framework for Quantitative Analysis of Machine-Learned Interatomic Model Training and Simulation Data

• PEGASUS: Unlocking Polarity in Cell-Permeable Cyclic Peptides Using AI Models Built on Massively Parallel Biological Assays

• Small Single-Stranded DNA Structure Prediction

• CaliciBoost: Performance-driven evaluation of molecular representations for caco-2 permeability prediction

• Multi-scale in-silico modelling to unveil structural requirements for DNA-PK inhibitors as radiosensitizers and MolSHAP based design of novel ligands

• Machine Learning-Guided Scope Selection to Balance Performance and Substrate Similarity

• The Impact of Reward Scalarization and Weight Scheduling on Optimization Dynamics in Multi-Objective Molecular Design

• Higher-level Strategies for Computer-Aided Retrosynthesis

• ASPEN: Robust detection of allelic dynamics in single cell RNA-seq

• RetroScore: graph edit distance-guided retrosynthesis for accessibility scoring with route metrics

• ProjFusNet: deep neural network for peptide precursor prediction using projection-fused protein language model and structural features

• qMol: A Web Server for Efficient Molecular Queries Using Fragment-Based Reduced Graphs

• Molecules in Wikipedia: Analysis of Their Chemical Diversity, Functional Roles, and Popularity

Archive

I’m quite interested in meaningful benchmarks. So, here are some standout papers on model validation and modellability I’ve read this year. This isn’t an exhaustive list, so please send over any of your own favorites that I missed.

• Data Set Modelability by QSAR

• Roughness of Molecular Property Landscapes and Its Impact on Modellability

• k Nearest Neighbors QSAR Modeling as a Variational Problem:  Theory and Applications

• Novel Variable Selection Quantitative Structure−Property Relationship Approach Based on the k-Nearest-Neighbor Principle

• Study of Data Set Modelability: Modelability, Rivality, and Weighted Modelability Indexes

• Regression Modelability Index: A New Index for Prediction of the Modelability of Data Sets in the Development of QSAR Regression Models

• Kernel Target Alignment Parameter: A New Modelability Measure for Regression Tasks

• Shallow Representation Learning via Kernel PCA Improves QSAR Modelability

• Experimental Error, Kurtosis, Activity Cliffs, and Methodology: What Limits the Predictivity of Quantitative Structure–Activity Relationship Models?

• Practically Significant Method Comparison Protocols for Machine Learning in Small Molecule Drug Discovery

• Comparative Studies on Some Metrics for External Validation of QSPR Models

• Assessing Model Fit by Cross-Validation

• Robust Cross-Validation of Linear Regression QSAR Models

• Criterion for Evaluating the Predictive Ability of Nonlinear Regression Models without Cross-Validation

• Comments on the Definition of the Q2 Parameter for QSAR Validation

• Evaluation of Cross-Validation Strategies in Sequence-Based Binding Prediction Using Deep Learning

• A Historical Excursus on the Statistical Validation Parameters for QSAR Models: A Clarification Concerning Metrics and Terminology

• Real External Predictivity of QSAR Models: How To Evaluate It? Comparison of Different Validation Criteria and Proposal of Using the Concordance Correlation Coefficient

• Real External Predictivity of QSAR Models. Part 2. New Intercomparable Thresholds for Different Validation Criteria and the Need for Scatter Plot Inspection

• Metric Validation and the Receptor-Relevant Subspace Concept

• Leave-Cluster-Out Cross-Validation Is Appropriate for Scoring Functions Derived from Diverse Protein Data Sets

• Comments on “Leave-Cluster-Out Cross-Validation Is Appropriate for Scoring Functions Derived from Diverse Protein Data Sets”: Significance for the Validation of Scoring Functions

• Statistical Confidence for Variable Selection in QSAR Models via Monte Carlo Cross-Validation

• Time-Split Cross-Validation as a Method for Estimating the Goodness of Prospective Prediction.

• External Validation and Prediction Employing the Predictive Squared Correlation Coefficient — Test Set Activity Mean vs Training Set Activity Mean

• Maximum Unbiased Validation (MUV) Data Sets for Virtual Screening Based on PubChem Bioactivity Data

• Practically Significant Method Comparison Protocols for Machine Learning in Small Molecule Drug Discovery

Palate Cleanser

Ash Jogalekar@curiouswavefn

"The reaction has an odd name too: the cyclization is sometimes called the Hajos–Parrish–Eder–Sauer–Wiechert reaction, after its discoverers, but only by those who want to impress their friends." - Clayden et al. Organic Chemistry

4:17 AM · Dec 30, 2025 · 1.13K Views

---

1 Repost · 9 Likes

Tommy Siegel@TommySiegel

I love that the New Yorker put this one in print this year, lol

3:03 PM · Dec 23, 2025 · 616K Views

---

63 Replies · 843 Reposts · 23.4K Likes

60% Water By Mass@MoleculesGoHard

People don't even ask for your PdCl2(PPh3)2 anymore they just look at you like

5:33 AM · Nov 18, 2025 · 1.84K Views

---

5 Reposts · 77 Likes

Lior Pachter@lpachter

Single-cell genomics finally makes it into the clinic. “Can you show me on the UMAP where it hurts?”

4:11 PM · Nov 16, 2025 · 53.6K Views

---

17 Replies · 94 Reposts · 845 Likes

---

---

Have a wonderful New Year,
Manas