A collection of papers (and code) that were at one time or another deemed interesting enough to hang on to…

Deep Learning

• Mixture of A Million Experts
  Xu Owen He
  ArXiv Preprint 2024

• Graph Neural Networks with Learnable Structural and Positional Representations + GitHub Repo
  Vijay Prakash Dwivedi, Anh Tuan Luu, Thomas Laurent, Yoshua Bengio, Xavier Bresson
  ICLR 2022

• Do Transformers Really Perform Badly for Graph Representation? + GitHub Repo
  Chengxuan Ying, Tianle Cai, Shengjie Luo, Shuxin Zheng, Guolin Ke, Di He, Yanming Shen, Tie-Yan Liu
  NeurIPS 2021
    Microsoft’s Graphormer paper

• Denoising Diffusion Probabilistic Models + GitHub Repo + Website
  Jonathan Ho, Ajay Jain, Pieter Abbee
  NeurIPS 2020

• Attention is All You Need + GitHub Repo (archived)
  Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin
  NeurIPS 2017
    The one, the only… Original transformer paper. This code eventually becomes the 🤗 Transformers library

• Adversarial Autoencoders - No official implementation available
  Alireza Makhzani, Jonathon Shlens, Navdeep Jaitly, Ian Goodfellow, and Brendan Frey
  ArXiv Preprint 2016

• PointerNets
  Oriol Vinyals, Meire Fortunato, and Navdeep Jaitly
  NeurIPS 2015

LLMs and Agents

• Instruction-Following Pruning for Large Language Models
  Bairu Hou, Qibin Chen, Jianyu Wang, Guoli Yin, Chong Wang, Nan Du, Ruoming Pang, Shiyu Chang, Tao Lei
  cs.CL on arXiv 2025
   The paper proposes “instruction-following pruning,” a dynamic structured pruning method for large language models (LLMs). It utilizes a sparse mask predictor that adapts based on user instructions, optimizing both the predictor and the LLM using instruction-following data. Results show that a 3B activated model outperforms a 3B dense model by 5-8 points in specific domains, matching a 9B model’s performance.

• Inconsistency of LLMs in Molecular Representations
  Bing Yan, Angelica Chen, Kyunghyun Cho
  Theoretical and Computational Chemistry on ChemRxiv 2024
   The paper investigates the consistency of large language models (LLMs) in molecular representations like SMILES and IUPAC names. Despite finetuning with a dual representation dataset and applying a Kullback-Leibler divergence loss for training, the models exhibited less than 1% consistency and no improvement in accuracy. Findings highlight the limitations of LLMs in understanding chemistry.

• MemGPT: Towards LLMs as Operating Systems + GitHub Repo
  Charles Packer, Sarah Wooders, Kevin Lin, Vivian Fang, Shishir G. Patil, Ion Stoica, and Joseph E. Gonzalez
  arXiv 2024
    Infinite context for lanuage models. Now pacakged as part of Letta.

• AvaTaR: Optimizing LLM Agents for Tool Usage via Contrastive Reasoning + GitHub Repo
  Shirley Wu, Shiyu Zhao, Qian Huang, Kexin Huang, Michihiro Yasunaga, Kaidi Cao, Vassilis N. Ioannidis, Karthik Subbian, Jure Leskovec, and James Zou
  NeurIPS 2024

• ReAct: Synergizing Reasoning and Acting in Language Models + Project Site
  Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik Narasimhan, and Yuan Cao
  ICLR 2023
    ReAct integrates reasoning and acting in LLMs by interleaving reasoning traces with task-specific actions. It reduces hallucinations in QA (HotpotQA) and fact verification (Fever) via Wikipedia API interactions and outperforms imitation and RL methods in ALFWorld (+34%) and WebShop (+10%). ReAct enhances interpretability and decision-making with minimal in-context examples.

• LoRA: Low-Rank Adaptation of Large Language Models + GitHub Repo
  Edward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen
  ICLR 2022
    LoRA is now wrapped into the 🤗 PEFT library

• Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
  Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, and Douwe Kiela
  NeurIPS 2020

Neural Reasoning & Decision Making

• DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning + Community Sourced GitHub Reproduction
  DeepSeek-AI,… Zhen Zhang
  cs.CL on arXiv 2025
    DeepSeek-R1 introduces reasoning-focused LLMs trained via reinforcement learning (RL). DeepSeek-R1-Zero, trained without supervised fine-tuning, shows strong reasoning but struggles with readability. DeepSeek-R1 improves performance using multi-stage training and cold-start data, matching OpenAI-o1-1217 on reasoning tasks. The study releases DeepSeek-R1-Zero, DeepSeek-R1, and six distilled models (1.5B–70B) based on Qwen and Llama.

• DeepSeek-V3 Technical Report + GitHub Repo
  DeepSeek-AI,… Zizheng Pan
  cs.CL on arXiv 2024
    DeepSeek-V3 is a 671B-parameter Mixture-of-Experts (MoE) model with 37B active parameters per token. It employs Multi-head Latent Attention (MLA) and DeepSeekMoE for efficiency and stability, trained on 14.8T tokens with SFT and RL. It outperforms open models, rivals closed ones, and trains efficiently in 2.788M H800 GPU hours without loss spikes.

• Forest-of-Thought: Scaling Test-Time Compute for Enhancing LLM Reasoning
  Zhenni Bi, Kai Han, Chuanjian Liu, Yehui Tang, Yunhe Wang
  arXiv 2024
    Forest-of-Thought (FoT) improves LLM reasoning by integrating multiple reasoning trees for collective decision-making. It employs sparse activation for efficiency, dynamic self-correction for real-time error handling, and consensus-guided decision-making. Experiments show FoT significantly enhances accuracy and efficiency in solving complex tasks, outperforming single-pass methods like Chain-of-Thought (CoT) and Tree-of-Thought (ToT).

• Quiet-STaR: Language Models Can Teach Themselves to Think Before Speaking
  Eric Zelikman, Georges Harik, Yijia Shao, Varuna Jayasiri, Nick Haber, and Noah D. Goodman
  CoLM 2024

• LLaVA-CoT: Let Vision Language Models Reason Step-by-Step + GitHub Repo
  Guowei Xu, Peng Jin, Hao Li, Yibing Song, Lichao Sun, and Li Yuan
  NeurIPS 2024

• Tree of Thoughts: Deliberate Problem Solving with Large Language Models + GitHub Repo
  Shunyu Yao, Dian Yu, Jeffrey Zhao, Izhak Shafran, Thomas L. Griffiths, Yuan Cao, Karthik Narasimhan
  NeurIPS 2023
    ReasoningAgent (Tree of Thought with Beam Search)

• STaR: Bootstrapping Reasoning With Reasoning
  Eric Zelikman, Yuhuai Wu, Jesse Mu, and Noah D. Goodman
  NeurIPS 2022

• Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
  Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Brian Ichter, Fei Xia, Ed Chi, Quoc Le, and Denny Zhou
  NeurIPS 2022

Contrastive Learning

• A Simple Framework for Contrastive Learning of Visual Representations + GitHub Repo
  Ting Chen, Simon Kornblith, Mohammad Norouzi & Geoffrey Hinton
  in Proceedings of the 37th International Conference on Machine Learning, PMLR 119:1597-1607, 2020

• FaceNet: A Unified Embedding for Face Recognition and Clustering
  Florian Schroff, Dmitry Kalenichenko, James Philbin
  in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015, 815

Recommender Systems

• A Scalable Recommendation Engine for New Users and Items
  Boya Xu, Yiting Deng, Carl Mela
  arXiv 2022

• Contextual Bandit Algorithms with Supervised Learning Guarantees
  Alina Beygelzimer, John Langford, Lihong Li, Lev Reyzin, Robert Schapire
  In Proceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics, PMLR 15:19-26, 2011

Molecular Generation

Generative models for molecules. Most typically text-based inputs (SMILES/SELFIES) or graph reps (parallel models on atom and bond matrices). Usually have some property optimization ability (latent space search/interpolation, reinformcement learning, guided genetic exploration). Most commonly these methods are autoregressive, but more recently non-autoregressive molecular generation methods have started to arise.

Reviews

• Machine learning-aided generative molecular design
  Yuanqi Du, Arian R. Jamasb, Jeff Guo, Tianfan Fu, Charles Harris, Yingheng Wang, Chenru Duan, Pietro Liò, Philippe Schwaller, and Tom L. Blundell
  Nat. Mach. Intell. 2024, 6, 589

• Deep Generative Models in De Novo Drug Molecule Generation
  Chao Pang, Jianbo Qiao, Xiangxiang Zeng, Quan Zou, and Leyi Wei
  J. Chem. Inf. Model. 2024, 64 (7), 2174

Diffusion / Flow Matching Models

• Exploring Discrete Flow Matching for 3D De Novo Molecule Generation + GitHub Repo
  Ian Dunn, and David R. Koes
  NeurIPS 2024
    Benchmarks discrete flow matching techniques for 3D molecular design. Paper introduces FlowMol-CTMC, a state-of-the-art model with fewer parameters than existing methods, demonstrating superior performance on small molecule generation. Other models often generate chemically valid but atypical functional groups outside the training distribution. Paper highlights gaps in structural motif prediction while advancing generative approaches for drug discovery.

• ShEPhERD: Diffusing shape, electrostatics, and pharmacophores for bioisosteric drug design + GitHub Repo (model) + GitHub Repo (scoring functions)
  Keir Adams, Kento Abeywardane, Jenna Fromer, and Connor W. Coley
  ArXiv 2024

• A dual diffusion model enables 3D molecule generation and lead optimization based on target pockets + GitHub Repo
  Lei Huang, Tingyang Xu, Yang Yu, Peilin Zhao, Xingjian Chen, Jing Han, Zhi Xie, Hailong Li, Wenge Zhong, Ka-Chun Wong & Hengtong Zhang
  Nat. Commun. 2024, 15, 2657

• Mixed Continuous and Categorical Flow Matching for 3D De Novo Molecule Generation + GitHub Repo
  Ian Dunn and David Ryan Koes
  ArXiv 2024
    Extends the flow matching framework to categorical data by constructing flows that are constrained to exist on a continuous representation of categorical data known as the probability simplex. Finds that, in practice, a simpler approach that makes no accommodations for the categorical nature of the data yields equivalent or superior performance. Presents FlowMol, a flow matching model for 3D de novo generative model that achieves improved performance over prior flow matching methods.

• GeoLDM: Geometric Latent Diffusion Models for 3D Molecule Generation + GitHub Repo
  Minkai Xu, Alexander Powers, Ron Dror, Stefano Ermon, and Jure Leskovec
  ICML 2023
    Stable (latent) diffusion model for 3D point clouds and 2D graphs. Capable of free and property conditioned generation (split-train-condition).

• Equivariant Diffusion for Molecule Generation in 3D + GitHub Repo
  Emiel Hoogeboom, Vı́ctor Garcia Satorras, Clément Vignac, and Max Welling
  in Proceedings of the 39th International Conference on Machine Learning, PMLR 162:8867-8887, 2022
    Non-autoregressive diffusion model (rotation invariant). Reps: \(x = (x_1 ... x_M) \in \mathbb{R}^{M \times 3}\) (atom position matrix) with corresponding feature vectors \(h = (h_1 ... h_M) \in \mathbb{R}^{M \times num feat}\).

Normalizing Flows

• MolGrow: A Graph Normalizing Flow for Hierarchical Molecular Generation (No implementation available)
  Maksim Kuznetsov and Daniil Polykovskiy
  Proceedings of the AAAI Conference on Artificial Intelligence 2021, 35 (9), 8226-8234
    Heirarchical normalizing flow for molecular graphs, autoregressive. Builds either BFS or fragment based (better). Model is composed of “plug-and-play” modules. Trained on MOSES, QM9, Zinc250k. Property-constrained optimization is based on genetic algorithm.

• FastFlows: Flow-Based Models for Molecular Graph Generation
  Nathan C. Frey, Vijay Gadepally, and Bharath Ramsundar
  ELLIS Machine Learning for Molecule Discovery Workshop 2021
    Framework for normalizing flows from SELFIES. Uses substructure filtering to speed up training and work from small training sets. Built in MPO functionality.
  TDS article

• MoFlow: An Invertible Flow Model for Generating Molecular Graphs + GitHub Repo
  Chengxi Zang and Fei Wang
  in Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining 2020
    Non-autoregressive normalizing Flow for molecular graphs; two-stage flow (bonds (based on GLOW network from Nvidia) > bond-conditioned flow for atoms). Similar to GraphNVP. Trained (NLL) on QM9 and ZINC250k. Developed new architecture. Excellent results.

• GraphAF: a Flow-based Autoregressive Model for Molecular Graph Generation + GitHub repo
  Chence Shi, Minkai Xu, Zhaocheng Zhu, Weinan Zhang, Ming Zhang, and Jian Tang
  ICLR 2020
    How to explain this better than reviewer #1…

"This paper proposes a generative model architecture for molecular graph generation based on autoregressive flows. The main contribution of this paper is to combine existing techniques (auto-regressive BFS-ordered generation of graphs, normalizing flows, dequantization by Gaussian noise, fine-tuning based on reinforcement learning for molecular property optimization, and validity constrained sampling). Most of these techniques are well-established either for data generation with normalizing flows or for molecular graph generation and the novelty lies in the combination of these building blocks into a framework."

GANs

• druGAN: An Advanced Generative Adversarial Autoencoder Model for de Novo Generation of New Molecules with Desired Molecular Properties in Silico - No official implementation available
  Artur Kadurin, Sergey Nikolenko, Kuzma Khrabrov, Alex Aliper, and Alex Zhavoronkov
  Mol. Pharmaceutics 2017, 14 (9), 3098–3104

• The cornucopia of meaningful leads: Applying deep adversarial autoencoders for new molecule development in oncology + GitHub Repo
  Artur Kadurin, Alexander Aliper, Andrey Kazennov, Polina Mamoshina, Quentin Vanhaelen, Kuzma Khrabrov, and Alex Zhavoronkov
  Oncotarget. 2017, 8, 10883-10890

• MolGAN: An implicit generative model for small molecular graphs + GitHub Repo
  Nicola De Cao and Thomas Kipf
  ICML 2018 Workshop on Theoretical Foundations and Applications of Deep Generative Models
    GAN for molecular graphs (tandem atom identity and bond matrices). Trained as W-GAN on QM9 with basic “RL” input. Implemented with R-GCNs.

Other

• Targeted Molecular Generation With Latent Reinforcement Learning
  Ragy Haddad, Eleni Litsa, Zhen Liu, Xin Yu, Daniel Burkhardt, Govinda Bhisetti
  ChemRxiv
  2025-01-03
   The paper presents a novel approach for targeted molecular generation using Reinforcement Learning with proximal policy optimization (PPO) in the latent space of pre-trained deep learning generative models. The method shows superior performance on benchmark datasets and can generate molecules with specific substructures while optimizing for desired properties, aiding drug discovery.

• TamGen: drug design with target-aware molecule generation through a chemical language model + GitHub Repo
  Kehan Wu, Yingce Xia, Pan Deng, Renhe Liu, Yuan Zhang, Han Guo, Yumeng Cui, Qizhi Pei, Lijun Wu, Shufang Xie, Si Chen, Xi Lu, Song Hu, Jinzhi Wu, Chi-Kin Chan, Shawn Chen, Liangliang Zhou, Nenghai Yu, Enhong Chen, Haiguang Liu, Jinjiang Guo, Tao Qin & Tie-Yan Liu
  Nat. Commun. 2024, 15, 9360

• TurboHopp: Accelerated Molecule Scaffold Hopping with Consistency Models
  Kiwoong Yoo, Owen Oertell, Junhyun Lee, Sanghoon Lee & Jaewoo Kang
  NeurIPS 2024

• DrugSynthMC: An Atom-Based Generation of Drug-like Molecules with Monte Carlo Search
  Milo Roucairol, Alexios Georgiou, Tristan Cazenave, Filippo Prischi & Olivier E. Pardo
  J. Chem. Inf. Model. 2024, 64, 18, 7097

• Enabling target-aware molecule generation to follow multi objectives with Pareto MCTS + GitHub Repo
  Yaodong Yang, Guangyong Chen, Jinpeng Li, Junyou Li, Odin Zhang, Xujun Zhang, Lanqing Li, Jianye Hao, Ercheng Wang & Pheng-Ann Heng
  Commun. Biol. 2024, 7, 1074

• Llamol: a dynamic multi-conditional generative transformer for de novo molecular design
  Niklas Dobberstein, Astrid Maass & Jan Hamaekers
  J. of Cheminf., 2024, 16, 73
    Transformer based on Llama2, tweaked for molgen. Not the most impressive paper, but some interesting tidbits scatted throughout (e.g., SCL, etc…)

• REINVENT4: Modern AI–driven generative molecule design + GitHub Repo
  Hannes H. Loeffler, Jiazhen He, Alessandro Tibo, Jon Paul Janet, Alexey Voronov, Lewis H. Mervin & Ola Engkvist
  J. of Cheminf., 2024, 16, 20
    AstraZeneca’s molecular design tool for de novo design, scaffold hopping, R-group replacement, linker design and molecule optimization.

• Masked graph modeling for molecule generation + GitHub Repo
  Omar Mahmood, Elman Mansimov, Richard Bonneau, and Kyunghyun Cho
  Nat. Commun. 2021, 12, 3156
    MPNN for moleular graphs. Generation by iterative sampling of subsets of graphs components, furuter generation steps are conditionalized on the rest of the graph. Trained on QM9 and ChEMBL. Paper provides analysis of GuacaMol benchmark metrics particularly their independence. Conclusions:

  1. Validity, KL-divergence and Fréchet Distance scores correlate highly with each other
  2. These three metrics correlate negatively with the novelty score
  3. Uniqueness does not correlate strongly with any other metric

Reaction Informatics

These models predict mechanisms for chemical reactions, ideally similar to how we teach 2nd years to push arrows. There are reltatively few of expamples of this task but they fall into 3 major categories electron flows, graph edits, reaction netowrks. At inference these models are used for forward synthesis prediction, potntially for prediction of chemo/regio-selectivity. Largely trained on pattern recognition from atom-mapped inputs (USPTO) though there are exceptions (e.g., Baldi papers below).

Electron Flow Prediction

• Non-Autoregressive Electron Redistribution Modeling for Reaction Prediction (NERF) + GitHub Repo
  Hangrui Bi, Hengyi Wang, Chence Shi, Connor Coley, Jian Tang, and Hongyu Guo
  in Proceedings of the 38th International Conference on Machine Learning, PMLR 139, 2021
    Models reactions as electron flow. Predicts flow non-autoregressively. Trained on USPTO-MIT; SOTA results. Graph rep from SMILES input. Fast (27x).

• Data-Efficient, Chemistry-Aware Machine Learning Predictions of Diels–Alder Reaction Outcomes + GitHub Repo
  Angus Keto, Taicheng Guo, Morgan Underdue, Thijs Stuyver, Connor W. Coley, Xiangliang Zhang, Elizabeth H. Krenske, and Olaf Wiest*
  J. Am. Chem. Soc., 2024, 146, 23, 16052–16061
    NERF applied to Diels-Alder cycloadditions. Typical training set reaches 90% accuracy (in line with general predictive accuracy…). No special improvements to NERF code released.

• A Generative Model For Electron Paths (ELECTRO) + GitHub Repo
  John Bradshaw, Matt J. Kusner, Brooks Paige, Marwin H. S. Segler, and José Miguel Hernández-Lobato
  ICLR 2019
    Learns to generates probabilbty distribution of electron paths. Trained on USPTO (w/ and w/o reaction condition data), 2e- chemistry only. Graph rep from SMILES input.

• Molecule Edit Graph Attention Network (MEGAN): Modeling Chemical Reactions as Sequences of Graph Edits + GitHub Repo
  Mikołaj Sacha, Mikołaj Błaż, Piotr Byrski, Paweł Dąbrowski-Tumański, Mikołaj Chromiński, Rafał Loska, Paweł Włodarczyk-Pruszyński, and Stanisław Jastrzębski
  J. Chem. Inf. Model. 2021, 61, (7), 3273–3284
    Not technically an electron flow model. Models chemical reations as series of graph edits, most similar to existing environment. Learns to predict sequences autoregressively.

Sources and Sinks

The Baldi papers map e- sources and sinks, combinatorially generates probability distribution of electron flows. Described classifiers are used to filter source-sink pairs before eval. Trained on in-house (unavailable) data. Papers don’t have available source code but ready-to-use programs are available on ChemDB.

• Deep learning for chemical reaction prediction
  David Fooshee, Aaron Mood, Eugene Gutman, Mohammadamin Tavakoli, Gregor Urban, Frances Liu, Nancy Huynh, David Van Vrankenb, and Pierre Baldi
  Mol. Syst. Des. Eng. 2018, 3, 442-452

• A Machine Learning Approach to Predict Chemical Reactions
  Matthew Kayala and Pierre Baldi
  NeurIPS 2011

• Learning to Predict Chemical Reactions
  Matthew A. Kayala, Chloé-Agathe Azencott, Jonathan H. Chen, and Pierre Baldi
  J. Chem. Inf. Model. 2011, 51 (9), 2209–2222

• PMechDB: A Public Database of Elementary Polar Reaction Step
  Mohammadamin Tavakoli, Ryan J. Miller, Mirana Claire Angel, Michael A. Pfeiffer, Eugene S. Gutman, Aaron D. Mood, David Van Vranken, and Pierre Baldi
  J. Chem. Inf. Model. 2024, 64, 6, 1975–1983
    Dataset paper for the Baldi papers. 100k polar elementary steps. PMechDB platform

Reaction Network Graphs

• Discovery of novel chemical reactions by deep generative recurrent neural network
  William Bort, Igor I. Baskin, Timur Gimadiev, Artem Mukanov, Ramil Nugmanov, Pavel Sidorov, Gilles Marcou, Dragos Horvath, Olga Klimchuk, Timur Madzhidov, and Alexandre Varnek
  Sci. Rep. 2021, 11, 3178

• Efficient prediction of reaction paths through molecular graph and reaction network analysis
  Yeonjoon Kim, Jin Woo Kim, Zeehyo Kim, and Woo Youn Kim
  Chem. Sci. 2018, 9, 825-835
    Search method for reaction intermediate networks. Uses DFT energies as heuristic.

Other

• Reaction rebalancing: a novel approach to curating reaction databases + GitHub Repo
  Tieu-Long Phan, Klaus Weinbauer, Thomas Gärtner, Daniel Merkle, Jakob L. Andersen, Rolf Fagerberg, and Peter F. Stadler
  J. Cheminf. 2024, 16, 82

• Transfer learning enables the molecular transformer to predict regio- and stereoselective reactions on carbohydrates + GitHub Repo
  Giorgio Pesciullesi, Philippe Schwaller, Teodoro Laino, and Jean-Louis Reymond
  Nat. Commun. 2020, 11, 4874
    Seq2Seq model for SMILES strings. Transfer learning allows for success on few-instance reactions. Blog post

Atom Mapping

• Precise atom-to-atom mapping for organic reactions via human-in-the-loop machine learning + GitHub Repo
  Shuan Chen, Sunggi An, Ramil Babazade, and Yousung Jung
  Nat. Commun. 2024, 15, 2250

• Bidirectional Graphormer for Reactivity Understanding: Neural Network Trained to Reaction Atom-to-Atom Mapping Task + GitHub Repo
  Ramil Nugmanov, Natalia Dyubankova, Andrey Gedich, and Joerg Kurt Wegner
  J. Chem. Inf. Model. 2022, 62, 14, 3307–3315
    Chython RxnMapper/Graphormer Mapper that’s heavily based on RXNMapper and CGR Tools. Semisupervised graph attention based model trained on USPTO and Pistachio datasets.

• Atom-to-atom Mapping: A Benchmarking Study of Popular Mapping Algorithms and Consensus Strategies
  Arkadii Lin, Natalia Dyubankova, Timur I. Madzhidov, Ramil I. Nugmanov, Jonas Verhoeven, Timur R. Gimadiev, Valentina A. Afonina, Zarina Ibragimova, Assima Rakhimbekova, Pavel Sidorov, Andrei Gedich, Rail Suleymanov, Ravil Mukhametgaleev, Joerg Wegner, Hugo Ceulemans, and Alexandre Varnek
  Mol. Inf. 2022, 41, 2100138

• Extraction of organic chemistry grammar from unsupervised learning of chemical reactions + GitHub Repo + RxnMapper app Webpage looks like it’s down
  Philippe Schwaller, Benjamin Hoover, Jean-Louis Reymond, Hendrik Strobelt, and Teodoro Laino
  Sci. Adv. 2021, 7 (15), eabe4166

• Reaction Data Curation I: Chemical Structures and Transformations Standardization + GitHub Repo
  Timur R. Gimadiev, Arkadii Lin, Valentina A. Afonina, Dinar Batyrshin, Ramil I. Nugmanov, Tagir Akhmetshin, Pavel Sidorov, Natalia Duybankova, Jonas Verhoeven, Joerg Wegner, Hugo Ceulemans, Andrey Gedich, Timur I. Madzhidov, and Alexandre Varnek
  Mol. Inf. 2021, 40, 2100119

Computer-Aided Retrosynthesis Planning

• Chimera: Accurate retrosynthesis prediction by ensembling models with diverse inductive biases
  Krzysztof Maziarz, Guoqing Liu, Hubert Misztela, Aleksei Kornev, Piotr Gaiński, Holger Hoefling, Mike Fortunato, Rishi Gupta, Marwin Segler
  arXiv 2024

• RLSynC: Offline–Online Reinforcement Learning for Synthon Completion
  Frazier N. Baker, Ziqi Chen, Daniel Adu-Ampratwum, and Xia Ning*
  J. Chim. Inf. Model. 2024, 64 (17), 6723

• Constrained synthesis planning with disconnection-aware transformer and multi-objective search
  Annie M. Westerlund, Lakshidaa Saigiridharan, Samuel Genheden
  ChemRxiv preprint 2024
    Retrosynthesis planning with the ability to constrain bonds, i.e., built in divergent synthesis constraints leading to shorter routes.
    AiZynthFinder with MO-MCTS and broken bonds score was used to run multistep experiments and can be found at: https://github.com/MolecularAI/aizynthfinder. Chemformer can be found at: https://github.com/MolecularAI/Chemformer. AiZynthTrain was used to tag disconnection-sites in the Chemformer training data and can be found at: https://github.com/MolecularAI/aizynthtrain.

• Retrosynthesis prediction using an end-to-end graph generative architecture for molecular graph editing + GitHub Repo
  Weihe Zhong, Ziduo Yang, and Calvin Yu-Chian Chen
  Nat. Commun. 2023, 14, 3009

Publication Parsing

• OpenChemIE: An Information Extraction Toolkit for Chemistry Literature + Web App
  Vincent Fan, Yujie Qian, Alex Wang, Amber Wang, Connor W. Coley, and Regina Barzilay
  J. Chem. Inf. Model. 2024, 64 (14), 5521

• Advancements in hand-drawn chemical structure recognition through an enhanced DECIMER architecture
  Kohulan Rajan, Henning Otto Brinkhaus, Achim Zielesny & Christoph Steinbeck
  J. Cheminf. 2024, 16, 76

ML Driven Drug Design

• MolE: a foundation model for molecular graphs using disentangled attention + GitHub Repo
  Oscar Méndez-Lucio, Christos A. Nicolaou & Berton Earnshaw
  Nat. Commun. 2024, 15, 9431
    Non-commercial license

• MolCompass: multi-tool for the navigation in chemical space and visual validation of QSAR/QSPR models + GitHub Repo
  Sergey Sosnin
  J. Cheminf. 2024, 16, 98

• HydraScreen: A Generalizable Structure-Based Deep Learning Approach to Drug Discovery + GitHub Repo
  Alvaro Prat, Hisham Abdel Aty, Orestis Bastas, Gintautas Kamuntavičius, Tanya Paquet, Povilas Norvaišas, Piero Gasparotto, and Roy Tal
  J. Chem. Inf. Model. 64 (15), 5817

• Exposing the Limitations of Molecular Machine Learning with Activity Cliffs
  Derek van Tilborg, Alisa Alenicheva, and Francesca Grisoni
  J. Chem. Inf. Model. 2022, 62 (23), 5938
    Overview of SAR cliffs and challenges for ML

Property/Activity Prediction

• Reusability report: exploring the utility of variational graph encoders for predicting molecular toxicity in drug design + GitHub Repo (NYAN) + GitHub Repo (Reuse) + GitHub Repo (Acute Tox MTL)
  Ruijiang Li, Jiang Lu, Ziyi Liu, Duoyun Yi, Mengxuan Wan, Yixin Zhang, Peng Zan, Song He & Xiaochen Bo
  Nat. Mach. Intell. 2024

• ChemXTree: A Feature-Enhanced Graph Neural Network-Neural Decision Tree Framework for ADMET Prediction
  Yuzhi Xu, Xinxin Liu, Wei Xia, Jiankai Ge, Cheng-Wei Ju, Haiping Zhang, John Z.H. Zhang
  J. Chem. Inf. Model. 2024, ASAP

• Quantitative structure–activity relationships of chemical bioactivity toward proteins associated with molecular initiating events of organ-specific toxicity + GitHub Repo
  Domenico Gadaleta, Marina Garcia de Lomana, Eva Serrano-Candelas, Rita Ortega-Vallbona, Rafael Gozalbes, Alessandra Roncaglioni & Emilio Benfenati
  J. Cheminformatics 2024, 16, 122

• A bioactivity foundation model using pairwise meta-learning + GitHub Repo
  Bin Feng, Zequn Liu, Nanlan Huang, Zhiping Xiao, Haomiao Zhang, Srbuhi Mirzoyan, Hanwen Xu, Jiaran Hao, Yinghui Xu, Ming Zhang & Sheng Wang
  Nat. Mach. Intell. 2024, 6, 962

• Ligand-Based Compound Activity Prediction via Few-Shot Learning + GitHub Repo
  Peter Eckmann, Jake Anderson, Rose Yu, and Michael K. Gilson
  J. Chem. Inf. Model. 2024, 64, (14), 5492

• QSARtuna: An Automated QSAR Modeling Platform for Molecular Property Prediction in Drug Design + GitHub Repo + Docs
  Lewis Mervin, Alexey Voronov, Mikhail Kabeshov, and Ola Engkvist
  J. Chem. Inf. Model. 2024, 64, (14), 5365

• Quantum-Informed Molecular Representation Learning Enhancing ADMET Property Prediction
  Jungwoo Kim, Woojae Chang, Hyunjun Ji, and InSuk Joung
  J. Chem. Inf. Model. 2024, 64, (13), 5028
    Supplementing a Graph Transformer with pretraing on DFT features. SoTA performance on 7 of 22 ADME-Tox tasks in TDC. Of particular interest is the data methods and architecture.

Active Learning Methods

• Human-in-the-loop active learning for goal-oriented molecule generation + GitHub Repo
  Yasmine Nahal, Janosch Menke, Julien Martinelli, Markus Heinonen, Mikhail Kabeshov, Jon Paul Janet, Eva Nittinger, Ola Engkvist & Samuel Kaski
  J. Cheminformatics 2024, 16, 138

• Finding the most potent compounds using active learning on molecular pairs + GitHub Repo
  Zachary Fralish, and Daniel Reker
  Beilstein J. Org. Chem. 2024, 20, 2152
    Extension of the matched pair methodology from the Reker group. Applied to Chemprop and XGB.

• DeepDelta: predicting ADMET improvements of molecular derivatives with deep learning + GitHub Repo
  Zachary Fralish, Ashley Chen, Paul Skaluba, and Daniel Reker
  J. Cheminformatics 2023, 15, 101
    Uses matched molecular pairs to predict property diverences. D-MPNN architecture is based on ChemProp, modified to take in 2 molecules.

Synthetic Accessibility

• Analog Accessibility Score (AAscore) for Rational Compound Selection + GitHub Repo
  Takato Ue, Akinori Sato, and Tomoyuki Miyao
  J. Chem. Inf. Model. 2024, ASAP

• Estimating the synthetic accessibility of molecules with building block and reaction-aware SAScore + GitHub Repo
  Shuan Chen and Yousung Jung
  J. Cheminf. 2024, 16, 83

Molecular Optimization

• A Zero-Shot Single-point Molecule Optimization Model: Mimicking Medicinal Chemists’ Expertise
  Peng Gao, Jie Zhang, Zhilian Dai, Yangyang Deng, Dan Zhang, Jiawei Fu, Songyou Zhong, Yichao Liu
  Theoretical and Computational Chemistry on ChemRxiv 2024
   The paper presents the Single-point Chemical Language Model (SpCLM), a framework for molecular design that mimics medicinal chemists’ expertise. Using a few hundred generated compounds, SpCLM predicts 60%-80% of active compounds in tests, correlating well with experimental data. This method reduces the need for extensive screening, offering a data-driven approach to optimize drug activity and selectivity.

• Projecting Molecules into Synthesizable Chemical Spaces
  Shitong Luo, Wenhao Gao, Zuofan Wu, Jian Peng, Connor W. Coley, and Jianzhu Ma
  ArXiv Preprint, 2024
    Interesting new approach to making molecules more synthesizable from genenerated virtual hits. Cleaning the chaff energy. Describes a new postfix notation (A B +) for synthetic transformations. Transformer-based model that translates graphs to postfix notation. Model capable of synthesis planning, generating similar and more synthesizable analogues, exploring chemical space in the syntesizablilty dimension.

• Evolutionary Multiobjective Molecule Optimization in an Implicit Chemical Space + GitHub Repo
  Xin Xia, Yiping Liu, Chunhou Zheng, Xingyi Zhang, Qingwen Wu, Xin Gao, Xiangxiang Zeng, and Yansen Su
  J. Chem. Inf. Model. 2024, 64, (13), 5161
    Multiobjective molecule optimization framework (MOMO) is a pareto-based MPO tool that evolves moelcules into better molecules. Genetic/ecolutionary algorithm in the latent (implicit) space ended by a VAE.

Large-scale Virtual Screening

• A novel bottom-up approach to find lead-compounds in billion-sized libraries
  Jordi Juárez-Jiménez, Xavier Barril, Álvaro Serrano-Morrás,… Carles Galdeano
  Theoretical and Computational Chemistry on ChemRxiv
   The paper presents a novel bottom-up approach for identifying lead compounds in vast chemical libraries, specifically trillion-scale on-demand collections. The method involves a systematic exploration of fragment space followed by focused mining of promising areas, utilizing advanced computational techniques to reduce false positives. This strategy successfully identified new BRD4 (BD1) binders with potency comparable to established drugs.

• Introducing SpaceGA: A Search Tool to Accelerate Large Virtual Screenings of Combinatorial Libraries + GitHub Repo
  Laust Moesgaard and Jacob Kongsted
  J. Chem. Inf. Model. 2024, 64 (21), 8123

• molli: A General Purpose Python Toolkit for Combinatorial Small Molecule Library Generation, Manipulation, and Feature Extraction + GitHub Repo
  Alexander S. Shved, Blake E. Ocampo, Elena S. Burlova, Casey L. Olen, N. Ian Rinehart & Scott E. Denmark
  J. Chem. Inf. Model. 2024, 64 (21), 8083

• PheSA: An Open-Source Tool for Pharmacophore-Enhanced Shape Alignment = GitHub Repo
  Joel Wahl
  J. Chem. Inf. Model. 2024, 64 (15), 5944-5953

• ROSHAMBO: Open-Source Molecular Alignment and 3D Similarity Scoring
  Rasha Atwi, Ye Wang, Simone Sciabola, and Adam Antoszewskim
  ChemRxiv Preprint 2024

• Pareto Optimization to Accelerate Multi-Objective Virtual Screening + GitHub Repo
  Jenna C. Fromer, David E. Graff, and Connor W. Coley
  arXiv preprint 2023
    Application of MolPAL (Molecular Pool-based Active Learning) to multi-objective virtual screening. Implements multiobjective Bayesian optimization to reduce the computational cost and apply it to the identification of ligands predicted to be selective based on docking scores to on- and off-targets. Identifies all pareto front molecules after 8% library exploration. Demonstrates superority of pareto optimization over scalarization.

• Accelerating high-throughput virtual screening through molecular pool-based active learning + GitHub Repo
  David E. Graff, Eugene I. Shakhnovicha, and Connor W. Coley
  Chem. Sci., 2021, 12, 7866
    Active learning tool for acceleration of virtual screening campaigns.

Cheminformatics

Reviews

• Modern chemical graph theory
  Leonardo S. G. Leite, Swarup Banerjee, Yihui Wei, Jackson Elowitt, Aurora E. Clark
  WIREs Comput. Mol. Sci. 2024, 14, (5), e1729

• Research Progresses and Applications of Knowledge Graph Embedding Technique in Chemistry
  Chuanghui Wang, Yunqing Yang, Jinshuai Song & Xiaofei Nan
  J. Chem. Inf. Model. 2024, 64, (19), 7213

General

• UNIQUE: A Framework for Uncertainty Quantification Benchmarking + GitHub Repo
  Jessica Lanini, Minh Tam Davide Huynh, Gaetano Scebba, Nadine Schneider, and Raquel Rodríguez-Pérez
  J. Chim. Inf. Model. 2024 ASAP
    Combo platform for model-agnostic uncertaintly quantification. Quasi-review paper.

• Topological Similarity Search in Large Combinatorial Fragment Spaces
  Louis Bellmann, Patrick Penner, Matthias Rarey
  J. Chem. Inf. Model. 2021, 61, (1), 238

• Open-Source Approach to GPU-Accelerated Substructure Search + GitHub Repo
  Andrew J. Whitehouse, Melchor Sanchez-Martinez, Seyedeh Maryam Salehi, Natalja Kurbatova, and Euan Dean
  J. Chem. Inf. Model. 2024, 64, (18), 6993

• When Do Quantum Mechanical Descriptors Help Graph Neural Networks to Predict Chemical Properties?
  Shih-Cheng Li, Haoyang Wu, Angiras Menon, Kevin A. Spiekermann, Yi-Pei Li & William H. Green
  J. Am. Chem. Soc. 2024, 146, (33), 23103

• Hilbert-curve assisted structure embedding method + GitHub Repo
  Gergely Zahoránszky-Kőhalmi, Kanny K. Wan & Alexander G. Godfrey
  J. Cheminf. 2024, 16, 87

Δ-machine learning

• Combining Hammett σ constants for Δ-machine learning and catalyst discovery + GitHub Repo
  V. Diana Rakotonirina, Marco Bragato, Stefan Heinenc & O. Anatole von Lilienfeld
  Digital Discovery 2024, ASAP

• Big Data meets Quantum Chemistry Approximations: The Δ-Machine Learning Approach
  Raghunathan Ramakrishnan, Pavlo O. Dral, Matthias Rupp, O. Anatole von Lilienfeld
  J. Chem. Theory Comput. 2015, 11, (5), 2087

Protein Structure Prediction

• Accurate structure prediction of biomolecular interactions with AlphaFold 3 - No code released
  Josh Abramson, Jonas Adler, Jack Dunger, … & John M. Jumper
  Nat. 2024 630, 493
    AlphaFold 3, a diffusion-based architecture that is capable of predicting the joint structure of complexes including proteins, nucleic acids, small molecules, ions and modified residues. Blog at Isomorphic

• State-specific protein-ligand complex structure prediction with a multiscale deep generative model + GitHub Repo
  Zhuoran Qiao, Weili Nie, Arash Vahdat, Thomas F. Miller III & Animashree Anandkumar
  Nat. Mach. Intell. 2024, 6, 195
    NeuralPLexer, a computational approach that can directly predict protein-ligand complex structures solely using protein sequence and ligand molecular graph inputs. Owing to its specificity in sampling both ligand-free-state and ligand-bound-state ensembles, NeuralPLexer consistently outperforms AlphaFold2 in terms of global protein structure accuracy on both representative structure pairs with large conformational changes and recently determined ligand-binding proteins. Some interesting takes in the OpenReview for their ICLR 2023 submission.
    Very interesting follow up study/comment from Pat Walters about test set infiltration and data purity (see LinkedIn for some spicy debate about this preprint).

• DiffDock: Diffusion Steps, Twists, and Turns for Molecular Docking + GitHub Repo
  Gabriele Corso, Hannes Stärk, Bowen Jing, Regina Barzilay, and Tommi Jaakkola
  ICLR 2023
    Cool paper that treats docking as a generative task instead of a search/regression task. DiffDock is a diffusion model over the non-Euclidean manifold of ligand poses. Really interesting way of thinking of things.

• Structure-based Drug Design with Equivariant Diffusion Models (DiffSBDD) + GitHub Repo
  Arne Schneuing, Yuanqi Du, Charles Harris, Arian Jamasb, Ilia Igashov, Weitao Du, Tom Blundell, Pietro Lió, Carla Gomes, Max Welling, Michael Bronstein, and Bruno Correia
  ArXiv Preprint 2022
    Diffusion model for SBDD, serious issues with results in this paper see OpenReview

• Accelerating Inference in Molecular Diffusion Models with Latent Representations of Protein Structure + GitHub Repo
  Ian Dunn and David Ryan Koes
  NeurIPS 2023
    GNN-based architecture for learning latent representations of molecular structure. Encodes protein represntation into reduced set of key points. When trained end-to-end with a diffusion model (DiffSBDD) for de novo ligand design, achieves comparable performance to one with an all-atom protein representation while exhibiting a 3-fold reduction in inference time. Unclear whether or not the original issues with DiffSBDD were address in this implementation…

Chemistry

Med Chem

• Escape from Flatland: Increasing Saturation as an Approach to Improving Clinical Success
  Frank Lovering, Jack Bikker, and Christine Humblet
  J. Med. Chem. 2009, 52 (21), 6752-6756

• Escape from Flatland 2: complexity and promiscuity
  Frank Lovering
  Med. Chem. Commun. 2013, 4, 515-519

My papers

• Sodium Alkyl(trimethylsilyl)amides: Substituent- and Solvent-dependent Solution Structures and Reactivities
  Qiulin You, Yun Ma, Ryan A. Woltornist, Nathan M. Lui, Jesse A. Spivey, Ivan Keresztes, David B. Collum
  Journal of the American Chemical Society 2024, 146 (44), 30397

• Natural Product Isolation of the Extract of Cleome rupicola Fruits Exhibiting Antioxidant Activity
  Yumi Gambrill, Patrick Commins, Stefan Schramm, Nathan M. Lui, Shaikha S. AlNeyadi, Panče Naumov
  Chemistry & Biodiversity 2024, 21 (4), e20230138

• Structure-Selectivity Principles Underlying Alkylations of Oppolzer’s Camphorsultam Enolates
  Nathan M. Lui
  Cornell University ProQuest Dissertations & Theses 2023, 30570998

• Sodiated Oppolzer enolates: solution structures, mechanism of alkylation, and origin of stereoselectivity
  Nathan M. Lui and David B. Collum
  Organic Chemistry Frontiers 2023, 10 (19), 4750

• MoFlowGAN: Combining adversarial and likelihood learning for targeted molecular generation + GitHub Repo
  Nathan M. Lui, Max D. Li, and Matt Ford
  ChemRxiv Preprint 2023

• Lithiated Oppolzer Enolates: Solution Structures, Mechanism of Alkylation, and the Origin of Stereoselectivity
  Nathan M. Lui, Samantha N. MacMillan, and David B. Collum
  Journal of the American Chemical Society 2022, 144 (51), 23379

• Sodium Isopropyl(trimethylsilyl)amide (NaPTA): A Stable and Highly Soluble Lithium Diisopropylamide Mimic
  Yun Ma, Nathan M. Lui, Ivan Keresztes, Ryan A. Woltornist, and David B. Collum
  The Journal of Organic Chemistry 2022, 87 (21), 14223

• Spectrochemistry of firefly bioluminescence
  Marieh B. Al-Handawi, Srujana Polavaram, Anastasiya Kurlevskaya, Patrick Commins, Stefan Schramm, César Carrasco-López, Nathan M. Lui, Kyril M. Solntsev, Sergey P. Laptenok, Isabelle Navizet, and Panče Naumov
  Chemical Reviews 2022, 122 (16), 13207

• The elusive relationship between structure and colour emission in beetle luciferases
  César Carrasco-López, Nathan M. Lui, Stefan Schramm, and Panče Naumov
  Nature Reviews Chemistry 2021, 5 (1), 4

• Thermochemiluminescent peroxide crystals
  Stefan Schramm, Durga Prasad Karothu, Nathan M. Lui, Patrick Commins, Ejaz Ahmed, Luca Catalano, Liang Li, James Weston, Taro Moriwaki, Kyril M. Solntsev, and Panče Naumov
  Nature Communications 2019, 10 (1), 997

• pH-Dependent fluorescence from firefly oxyluciferin in agarose thin films
  Nathan M. Lui, Stefan Schramm, and Panče Naumov
  New Journal of Chemistry 2019, 43 (3), 1122

• Beetle luciferases with naturally red-and blue-shifted emission
  César Carrasco-López, Juliana C Ferreira, Nathan M. Lui, Stefan Schramm, Romain Berraud-Pache, Isabelle Navizet, Santosh Panjikar, Panče Naumov, Wael M. Rabeh
  Life Science Alliance 2018, 1 (4), e201800072