Preprint scanner project / Issue 2026-09-30

Useful Chemistry

A weekly review of arXiv, ChemRxiv, and bioRxiv for ai-enabled methodological advances. Up to twenty papers reviewed every week; a quiet week may publish none.

20 selected / 178 reviewed Sep 23, 2026–Sep 29, 2026
Groups
Group roll call
96 followed groups Research-group roll call
  • Milad Abolhasani Abolhasani Lab North Carolina State University Site ↗
  • Surl-Hee (Shirley) Ahn Ahn Lab University of California, Davis Site ↗
  • Mohammed AlQuraishi AlQuraishi Laboratory Columbia University Site ↗
  • Sayan Banerjee Banerjee Group University of Tennessee, Knoxville Site ↗
  • Christopher J. Bartel Bartel Research Group University of Minnesota Twin Cities Site ↗
  • Jan-Niklas Boyn Boyn Research Group University of Minnesota Twin Cities Site ↗
  • Ting Cao Cao Group University of Washington Site ↗
  • Timothy Cernak Cernak Lab University of Michigan Site ↗
  • Ming Chen Ming Chen Group Purdue University Site ↗
  • Po-Yen Chen Chen Research Group University of Maryland, College Park Site ↗
  • Bingqing Cheng Cheng Group University of California, Berkeley Site ↗
  • Brian Cleary Algorithmic Lens on Experimental Biology Laboratory Boston University Site ↗
  • Connor W. Coley Coley Research Group Massachusetts Institute of Technology Site ↗
  • Yamil J. Colón Colón Group University of Notre Dame Site ↗
  • César de la Fuente Machine Biology Group University of Pennsylvania Site ↗
  • Julia Dshemuchadse Cape Crystal Cornell University Site ↗
  • Thomas P. Fay Fay Group University of California, Los Angeles Site ↗
  • Kara D. Fong Fong Lab California Institute of Technology Site ↗
  • Thomas E. Gartner III Gartner Group Lehigh University Site ↗
  • Rafael Gómez-Bombarelli Learning Matter Massachusetts Institute of Technology Site ↗
  • Jonathan Gootenberg Gootenberg Laboratory Harvard University Site ↗
  • Prashun Gorai 3D Materials Lab Rensselaer Polytechnic Institute Site ↗
  • Diptarka Hait Hait Group Columbia University Site ↗
  • Brian Hie Laboratory of Evolutionary Design Stanford University Site ↗
  • Guoxiang Hu Hu Group Georgia Institute of Technology Site ↗
  • Yong-Jie Hu Materials Computation and Informatics Group Drexel University Site ↗
  • Yifei Huang Huang Lab Pennsylvania State University Site ↗
  • Yunha Hwang Hwang Lab Massachusetts Institute of Technology Site ↗
  • Nicholas E. Jackson AI for Materials Group University of Illinois Urbana-Champaign Site ↗
  • William M. Jacobs Jacobs Group Princeton University Site ↗
  • Dipti Jasrasaria Jasrasaria Group University of Chicago Site ↗
  • Anupama Jha Jha Lab Yale University Site ↗
  • Kaiyi Jiang Jiang Lab Princeton University Site ↗
  • Adrian Jinich Jinich Lab University of California, San Diego Site ↗
  • Felipe Jornada Jornada Research Group Stanford University Site ↗
  • Kalli Kappel Kappel Lab University of California, Los Angeles Site ↗
  • Joshua Kretchmer Kretchmer Research Group Georgia Institute of Technology Site ↗
  • Aditi S. Krishnapriyan Krishnapriyan Research Group University of California, Berkeley Site ↗
  • Sebastian Kube Kube Lab University of Wisconsin–Madison Site ↗
  • Heather J. Kulik Kulik Research Group Massachusetts Institute of Technology Site ↗
  • Ambarish R. Kulkarni Kulkarni Research Group University of California, Davis Site ↗
  • Joseph S. Kwon Kwon Research Group The Ohio State University Site ↗
  • Joonho Lee Lee Group Harvard University Site ↗
  • Can Li Li Research Group Purdue University Site ↗
  • Wanlu Li Wanlu Li Research Group University of California, San Diego Site ↗
  • Rebecca K. Lindsey Lindsey Lab University of Michigan Site ↗
  • Ge Liu Ge Liu Group University of Illinois Urbana-Champaign Site ↗
  • Yuanyue Liu Yuanyue Liu Group The University of Texas at Austin Site ↗
  • Yang Lu Lu Lab University of Wisconsin–Madison Site ↗
  • Jiankun Lyu Evnin Family Laboratory of Computational Molecular Discovery The Rockefeller University Site ↗
  • Cong Ma Cong Ma Lab University of Michigan Site ↗
  • Arkajit Mandal Mandal Group Texas A&M University Site ↗
  • Andrew J. Medford Medford Research Group Georgia Institute of Technology Site ↗
  • Ilias Mitrai Systems and AI Lab The University of Texas at Austin Site ↗
  • Jeetain Mittal Mittal Group Texas A&M University Site ↗
  • Joel A. Paulson Paulson Lab University of Wisconsin–Madison Site ↗
  • Elisa Pieri Pieri Lab University of North Carolina at Chapel Hill Site ↗
  • Doran Raccah MesoScience Lab The University of Texas at Austin Site ↗
  • Phillip Rauscher Rauscher Group New York University Site ↗
  • Wesley Reinhart Reinhart Group Pennsylvania State University Site ↗
  • Gabriel J. Rocklin Rocklin Lab Northwestern University Site ↗
  • Andrew S. Rosen Rosen Research Group Princeton University Site ↗
  • Grant M. Rotskoff Rotskoff Group Stanford University Site ↗
  • Janani Sampath Sampath Research Group University of Florida Site ↗
  • Elvira Sayfutyarova Sayfutyarova Group Pennsylvania State University Site ↗
  • Martin Seifrid Seifrid Group North Carolina State University Site ↗
  • Thomas P. Senftle Senftle Group Rice University Site ↗
  • Karthik Shekhar Shekhar Lab University of California, Berkeley Site ↗
  • Zachary M. Sherman Z Lab University of Washington Site ↗
  • Krishna Shrinivas Shrinivas Lab Northwestern University Site ↗
  • Rohit Singh Singh Lab Duke University Site ↗
  • Micheline Soley Soley Group University of Wisconsin–Madison Site ↗
  • Kevin V. Solomon Solomon Laboratory University of Delaware Site ↗
  • Jeff Spence Spence Lab University of California, San Francisco Site ↗
  • Kayla G. Sprenger Rational Design of Interfaces Lab University of Colorado Boulder Site ↗
  • Chong Sun Sun Lab Rutgers University–New Brunswick Site ↗
  • Yidan Sun Sun Lab Washington University in St. Louis Site ↗
  • Daniel Tabor Tabor Research Group Texas A&M University Site ↗
  • Ming Tang Mesoscale Materials Science Group Rice University Site ↗
  • Roel Tempelaar Tempelaar Team Northwestern University Site ↗
  • Erik Thiede Thiede Lab Cornell University Site ↗
  • Pratyush Tiwary Artificial Chemical Intelligence@Maryland University of Maryland, College Park Site ↗
  • Brian Trippe Trippe Lab Stanford University Site ↗
  • Alexander Urban Urban Research Group Columbia University Site ↗
  • David Van Valen Van Valen Lab California Institute of Technology Site ↗
  • Vojtech Vlcek Vlcek Group University of California, Santa Barbara Site ↗
  • Allon Wagner Wagner Lab University of California, Berkeley Site ↗
  • Shunzhi Wang Wang Lab New York University Site ↗
  • Michael A. Webb Webb Research Group Princeton University Site ↗
  • Mingjian Wen Wen Research Group University of Houston Site ↗
  • Hong-Zhou Ye Ye Group University of Maryland, College Park Site ↗
  • Shuwen Yue Yue Research Group Cornell University Site ↗
  • Daiwei (David) Zhang Daiwei Zhang Lab University of North Carolina at Chapel Hill Site ↗
  • Hongbo Zhao Zhao Research Group University of California, San Diego Site ↗
  • Jian Zhou Zhou Lab University of Chicago Site ↗
  • Tianyu Zhu Zhu Group Yale University Site ↗

Issue 2026-09-30

Highlights from this week

#01 arXiv v1 / preprint

Preprint—not peer reviewed

Deep Learning GW Quasiparticle Hamiltonians for Many-Body Excited-State Electronic Structure at Scale

Xiaoxun Gong, Zechen Tang, Woochang Kim, Yang Li, Wenhui Duan, Yong Xu, Steven G. Louie

The authors introduce DeepH-GW to predict an effective GW quasiparticle Hamiltonian directly from atomic structure. They report quasiparticle band structures with errors on the order of a few meV and transferability to substantially larger supercells. This may provide a practical route toward large-scale many-body simulations.

AI–materials Electronic structureSurrogate modeling
Abstract brief CC0 Preprint
#02 bioRxiv v2 / preprint

Preprint—not peer reviewed

De novo design of flexible protein interactions with GuideFlip

Yi, K., Chen, Q., Zhang, D., Tian, P., Wagstaff, J. L., McLaughlin, S. H., Tate, C. G., Jamali, K., Scheres, S. H. W.

The authors progressively assign binder residues while re-predicting the complex at each step. They report de novo binders to the C-terminus of -synuclein and the disordered amino terminus of RBX1, with epitopes confirmed by NMR and mutagenesis. This could help design binders when bound structures emerge only upon binding.

AI–biochemistry Biomolecular designProtein engineering
Abstract brief CC BY Preprint
#03 arXiv v1 / preprint

Preprint—not peer reviewed

Linear-Scaling Quantum Transport from Machine-Learning Density Functional Theory Hamiltonians

Bang Liu, Yang Zhong, Zhi-Xin Guo, Xin-Gao Gong, Hongjun Xiang

The authors use a graph neural network trained on small-structure DFT Hamiltonians to predict Hamiltonians for large devices. They report faithful transmission spectra and predictive simulations beyond 10,000 atoms. This could make device-scale quantum-transport calculations more practical.

AI–materials Atomistic modelingElectronic structure
Abstract brief CC0 Preprint
#04 bioRxiv v1 / preprint

Preprint—not peer reviewed

MIMOSA: Guiding De Novo Antibody Design with Natural Interaction Fingerprints

Abanades, B., Roncoli, A., Bhagawati, M., Kessel, P., Imhof-Jung, S., Doerr, D., Schilz, J., Vasilaki, S., Seeger, F., Bonvin, A. M. J. J., Bonneau, R., Gligorijevic, V., Vangone, A.

The authors constrain antibody diffusion designs to retain the geometry and chemistry of a known binding interface. They report experimentally validated binders at hit rates from 20 to 26 percent across three targets where unconstrained methods yielded few or none. Using an existing interface as a design constraint could expand de novo antibody work to more difficult targets.

AI–biochemistry Biomolecular designGenerative design
Abstract brief CC BY Preprint
#05 arXiv v1 / preprint

Preprint—not peer reviewed

Reinforcement learning amortizes transition-state physics into one-step flow models

Yunyang Li, Zechang Sun, Kuang Yu, Wen Yan, Mark Gerstein, Hung Q. Pham

The authors introduce Trèfle, a one-step flow model that moves physics-guided generation into training with a mean-flow objective, a low-cost surrogate, and reinforcement-learning post-training. They report accurate transition states at 30× the throughput of multi-step baselines. This reported speed could support automated reaction discovery.

AI–chemistry Generative designReaction modeling
Abstract brief CC0 Preprint
#06 bioRxiv v1 / preprint

Preprint—not peer reviewed

RNASeek: A Cross-Phyla Generative Foundation Model for Multipurpose RNA Modeling and Reinforcement Learning-Based Design

Chen, S., Fernandes, N., Li, W. V., Wang, L., Jia, Z., Xiang, J. S.

The authors use functional predictors as reward models to update RNASeek toward sequences with desired properties. They report experimentally validated ribozymes with wild-type activity and 3' UTRs that exceeded training data and benchmarked AI-generated sequences. This may connect learned RNA function to controllable de novo sequence design.

AI–biochemistry Biomolecular designFoundation modelsGenerative design
Abstract brief CC BY Preprint
#07 arXiv v2 / preprint

Preprint—not peer reviewed

AI-guided high-throughput discovery of iridium- and ruthenium-free palladium-oxide catalysts for durable acidic oxygen evolution

Ken J. Jenewein, Faezeh Habib Zadeh, Xiaoxiao Wang, Gustavo Malkomes, Huafan Zhang, Natalie Page, Jae Jin Bang, Peter J. Santiago, Karla V. Contreras, Katherine K. Li, Allison Perna, Lorena M. Britton, Fahrettin Kilic, Kevin J. Cruse, Armin Taheri, Krishnanand Mallayya, Harley Quinn, Rebecca A. Durr, Peter A. Beaucage, Santiago Miret, John M. Gregoire, Rafael Gomez-Bombarelli

The authors report an AI-guided, human-supervised closed-loop platform combining sputter synthesis, high-throughput screening, composition-property models, and adaptive optimization. They report that InMnPdOx retained an overpotential below 0.5 V over 1,000 h of operation. The result suggests a way to pursue activity and stability together.

AI–chemistryAI–materials Autonomous labsGenerative design
Abstract brief CC0 Preprint
#08 arXiv v1 / preprint

Preprint—not peer reviewed

Combining physical models with dynamically acquired experimental information for the optimization of multicomponent NASICON fast ionic conductors in a self-driving laboratory

Bernardus Rendy, Yuxing Fei, Tanjin He, Xiaochen Yang, Andrea Giunto, Lauren N. Walters, David Milsted, Hao Qiu, Matthew J. McDermott, Bin Ouyang, Yan Zeng, Gerbrand Ceder

The authors use aggregated cost functions to jointly guide autonomous experiments by synthetic accessibility and target properties. In 78 autonomous trials, they report 18 promising compositions and two fast-conducting solid electrolytes. Coupling failed and successful syntheses to the design model could make materials optimization more informative as well as more efficient.

AI–materials Autonomous labsGenerative design
Abstract brief CC0 Preprint
#09 arXiv v1 / preprint

Preprint—not peer reviewed

Bridging Molecular Scales with Implicit Score Matching for Bottom-Up Coarse Graining

Patrick G. Sahrmann, Nicholas Lubbers, Benjamin T. Nebgen, Emily Shinkle, Sergei Tretiak, Kipton Barros,, Brenden W. Hamilton

The authors use implicit score matching to avoid explicit coarse-grained forces. They report high-fidelity models in which each coarse-grained site represents hundreds to thousands of atoms. This could help bridge molecular and mesoscopic simulation scales.

AI–chemistryAI–materials Atomistic modelingMultiscale modeling
Abstract brief CC0 Preprint
#10 bioRxiv v1 / preprint

Preprint—not peer reviewed

AtomWeaver: Multi-Component Flow Matching with a Structured Geometric Prior Facilitates Non-Canonical Peptide Design

Kitaygorodsky, A., Hostallero, D. E., Broom, A., Layne, E., Hwang, S., Kanawaty, A. K., Babej, T., Butterfoss, G. L., Fingerhuth, M.

The authors generate peptide side-chain atoms as unlabeled points and assign identity by matching predicted clouds to a template library. They report broad coverage of a 300-residue vocabulary, including four non-canonical types not visible in training. Choosing identity at decoding may make peptide design adaptable to new residue libraries without retraining.

AI–biochemistry Biomolecular designGenerative design
Abstract brief CC BY Preprint
#11 arXiv v1 / preprint

Preprint—not peer reviewed

Symmetry-Aware Flow Matching for End-to-end Molecular Crystal Generation

Wendi Cai, Fanyang Mo

The authors evolve asymmetric-unit coordinates and symmetry-compatible lattice parameters while using the full periodic environment for updates. Across held-out crystals, they report a 72.0% hit rate from 50 candidates per target, versus 8.1% for a full-cell baseline. This separation may improve symmetry-constrained molecular-crystal generation.

AI–chemistryAI–materials Generative designMaterials representation
Abstract brief CC0 Preprint
#12 arXiv v1 / preprint

Preprint—not peer reviewed

A System-Independent Metadynamics Strategy for Reactive Training Data: Application to Gas-Phase Organic Reactions

Wanrun Jiang, Jinzhe Zeng, Manyi Yang, Tong Zhu, Han Wang

The authors introduce a system-independent collective variable based on Cartesian RMSD in randomly partitioned local domains and time-averaged reference geometries to drive metadynamics. They report OpenRxn26, a dataset of 1.8 M DFT-labeled configurations for neutral singlet unimolecular reactions in H/C/N/O chemical space. These results suggest that broader sampling could help produce reactive training data for dynamics.

AI–chemistry Datasets + benchmarksNeural potentialsReaction modeling
Abstract brief CC0 Preprint
#13 arXiv v1 / preprint

Preprint—not peer reviewed

One Sequence, Many Decodings: CAGenMol-2 Recasts Drug Design as Masked Molecular Inference

Yanting Li, Enyan Dai, Lei Wang, Wen-Cai Ye, Li Liu

The authors select property prediction, conditional generation, or partial-constraint design by masking regions of a molecular sequence. On CrossDocked2020, they report a Success Rate increase from 30.2% to 70.8% while largely preserving drug-likeness and diversity. A shared sequence interface could reduce the need for separate models across compound design tasks.

AI–chemistry Foundation modelsGenerative design
Abstract brief CC0 Preprint
#14 arXiv v1 / preprint

Preprint—not peer reviewed

Graph Forward Distribution Matching for Molecular Inverse Design

Yihan Zhu, Yuhan Liu, Brett Savoie, Tengfei Luo, Meng Jiang

The authors optimize graph diffusion through its forward process using a reward-tilted target distribution. For multi-conditional polymer and small-molecule generation, they report the lowest MAE on every target property and chemical validity above 0.99. This may offer a route to multi-property molecular design while retaining valid graphs.

AI–chemistryAI–materials Generative design
Abstract brief CC0 Preprint
#15 arXiv v1 / preprint

Preprint—not peer reviewed

Atom-Resolved Machine Learning of Dielectric and Piezoelectric Response

Jinyu Liu, Yingwei Chen, Liyang Ma, Hongyu Yu, Hongjun Xiang

The authors learn ionic response as an atom-indexed field and reconstruct material response through sparse linear solves. They report identifying a polar superlattice candidate whose measured response exceeded the ordered reference by 63 and 66 percent. Atom-resolved response models could help screen structurally complex dielectric and piezoelectric materials without the full cost of direct calculations.

AI–materials Electronic structureSurrogate modeling
Abstract brief CC0 Preprint
#16 arXiv v1 / preprint

Preprint—not peer reviewed

Curvature-Aware Flow Matching for Molecular Structure Generation

Samir Darouich, Juliane Geisler, Jacob W. Toney, Tanja Bien, Johannes Kastner, Heather J. Kulik, Mathias Niepert

The authors add local energy-surface curvature to flow-model training through a frozen learned Hessian predictor. They report a 63 percent reduction in median maximum-force deviation for conformers and as much as a 10 percent rise in valid transition states. Curvature supervision may improve physical plausibility where coordinate accuracy alone leaves the local landscape underspecified.

AI–chemistry Generative designReaction modeling
Abstract brief CC0 Preprint
#17 bioRxiv v1 / preprint

Preprint—not peer reviewed

GEMOT: Towards Mechanistic World Models for Biology

de Pomereu, T., Snelling, B., Froehlich, F.

The authors record hypotheses in a semantic layer and separate hypothesis formulation from numerical evaluation. They report that autonomously constructed models matched or exceeded reference models in fit and parsimony across the evaluated problems. This audit trail may make automated mechanistic modeling easier to inspect.

AI–biochemistry Scientific agents
Abstract brief CC BY Preprint
#18 bioRxiv v1 / preprint

Preprint—not peer reviewed

Quantitative machine learning of protein interactions reveals the multiscale organization and molecular syntax of signaling networks

Rogers, J. R., Cardani, L. P., Shah, N. H., AlQuraishi, M.

The authors use a biophysically informed neural network and data harmonization to predict domain-peptide dissociation constants from sequence. They report calibrated confidence and quantitative drafts of human signaling networks analyzed across three physical scales. Quantitative affinity estimates could provide a more mechanistic way to assemble and interrogate signaling-network models.

AI–biochemistry Surrogate modeling
Abstract brief CC BY Preprint
#19 arXiv v1 / preprint

Preprint—not peer reviewed

Where Should Physics Enter a Molecular Crystal Generator?

Haocheng Tang, Junmei Wang, Wengong Jin

The authors compare physics added during training, after training, and at inference for all-atom crystal generators. They report that post-training improved molecular validity and packing, whereas relaxation repaired clashes but made generation 6 to 26 times slower. The comparison suggests a practical division between learned physical preferences and targeted corrections at generation time.

AI–chemistryAI–materials Generative designNeural potentials
Abstract brief CC0 Preprint
#20 arXiv v1 / preprint

Preprint—not peer reviewed

GLASS: Global Latent Aggregation with Slot-based Set Decoding for Scalable All-Atom Crystal Generation

Hendrik Kra\ss, Seyed Mohamad Moosavi, Mathias Niepert

The authors encode structures in a permutation-invariant global latent space and use a learned-slot decoder to construct atoms in parallel. They report generating MOFs with up to 150 atoms per unit cell without conditioning on building blocks, topology, or composition. This may support generation of larger atomistic systems while exposing a validity--novelty tradeoff.

AI–materials Generative designMaterials representation
Abstract brief CC0 Preprint