1. Agentic Preparative Thin-Layer Chromatography System for Autonomous Purification1
PLANAR combines robotic plate handling, solvent prediction and an agentic controller for preparative thin-layer chromatography. Three workflow demonstrations produced NMR-verified isolated compounds, addressing purification as a bottleneck after synthesis. The system is supervised and uses explicit review gates; it does not establish unattended general autonomy. Our account is restricted to the publisher-deposited abstract.1
2. Enabling Data-Efficient Machine Learning through Multi-Target Training Set Design via Greedy Neighborhood Coverage2
MT-RaRF selects a shared set of training reactions for several prediction targets by covering overlapping chemical neighborhoods. Across asymmetric-catalysis datasets, the method reportedly preserves comparable prediction errors while using fewer experiments than independently selecting neighbors for each target. The benefit depends on local overlap and budget. Our account is restricted to the publisher-deposited abstract.2
3. Enzymatic Reaction Feasibility Classification Using Machine Learning Methods3
Researchers compare reaction representations and classifiers for enzymatic reaction feasibility, combining four representations in the ERFC consensus model. Tests distinguish database reactions from rule-generated negatives and include a five-step biosynthetic pathway. The study supports prioritizing predicted pathways for investigation, while classification against constructed negatives does not establish experimental feasibility. Our account is restricted to the publisher-deposited abstract.3
4. Glucose Dehydrogenase Engineered for the Noncanonical Coenzyme NMN Outperforms Its Natural NAD-Dependent Activity4
An unreviewed study engineers glucose dehydrogenase to use the smaller NMN coenzyme, with a variant outperforming the natural enzyme–coenzyme pairing in reported activity and catalytic efficiency. Coupling the variant to four partner enzymes demonstrates NMN-recycling cascades. This broadens options for cofactor-dependent biocatalysis. Our account is restricted to the repository-deposited abstract.4
5. Mechanistically Agnostic Aliphatic N–H Transmutations Enabled by Interrupted Nitrogen Deletion5
Interrupted nitrogen deletion converts heteroaryl-fused piperidines into intermediates that support three different skeletal replacements: nitrogen-to-carbonyl, nitrogen-to-oxygen and nitrogen-to-sulfur. Experiments and calculations relate productive intermediate formation to the energetic cost of disrupting aromaticity. The work expands routes between related ring systems for molecular synthesis. Our account is restricted to the publisher-deposited abstract.5
References
Wendi Cai; Boxuan Zhao; Yansong Yue; Changlin Liu; Chengchun Liu; Ying Cui; Zihao Zhou; Xiaohui Tian; Zhongchao Zhang; Zhen Yang; Jie Zhu; Fanyang Mo. Agentic Preparative Thin-Layer Chromatography System for Autonomous Purification. Journal of the American Chemical Society; 2026; Peer-reviewed journal article, first online. DOI: 10.1021/jacs.6c10800. Accessed 2026-09-29T16:12:30.631Z.
Source evidence and access
Original deposited abstract and posted/first-online metadata
Paraphrase of the inspected original abstract: PLANAR combines robotic plate handling, solvent prediction and an agentic controller for preparative thin-layer chromatography. Three workflow demonstrations produced NMR-verified isolated compounds, addressing purification as a bottleneck after synthesis. The system is supervised and uses explicit review gates; it does not establish unattended general autonomy.
Restricted to original publisher/repository-deposited abstract and metadata in Crossref; landing page HTTP403 and full text not inspected.
Jakob A. Meckes; Jiang Yi He; Jolene P. Reid. Enabling Data-Efficient Machine Learning through Multi-Target Training Set Design via Greedy Neighborhood Coverage. ACS Catalysis; 2026; Peer-reviewed journal article, first online. DOI: 10.1021/acscatal.6c04894. Accessed 2026-09-29T16:12:30.632Z.
Source evidence and access
Original deposited abstract and posted/first-online metadata
Paraphrase of the inspected original abstract: MT-RaRF selects a shared set of training reactions for several prediction targets by covering overlapping chemical neighborhoods. Across asymmetric-catalysis datasets, the method reportedly preserves comparable prediction errors while using fewer experiments than independently selecting neighbors for each target. The benefit depends on local overlap and budget.
Restricted to original publisher/repository-deposited abstract and metadata in Crossref; landing page HTTP403 and full text not inspected.
Xin Wang; Hongyan Yin; Yekai Shen; Yushan Zhu; Igor V. Tetko; Aixia Yan. Enzymatic Reaction Feasibility Classification Using Machine Learning Methods. Journal of Chemical Information and Modeling; 2026; Peer-reviewed journal article, first online. DOI: 10.1021/acs.jcim.6c02294. Accessed 2026-09-29T16:12:30.633Z.
Source evidence and access
Original deposited abstract and posted/first-online metadata
Paraphrase of the inspected original abstract: Researchers compare reaction representations and classifiers for enzymatic reaction feasibility, combining four representations in the ERFC consensus model. Tests distinguish database reactions from rule-generated negatives and include a five-step biosynthetic pathway. The study supports prioritizing predicted pathways for investigation, while classification against constructed negatives does not establish experimental feasibility.
Restricted to original publisher/repository-deposited abstract and metadata in Crossref; landing page HTTP403 and full text not inspected.
Zhihua Nie; Zhan Song; Zhenyu Zhai; Yi-Heng P. Job Zhang. Glucose Dehydrogenase Engineered for the Noncanonical Coenzyme NMN Outperforms Its Natural NAD-Dependent Activity. ChemRxiv; 2026; Unreviewed preprint, version 1. DOI: 10.26434/chemrxiv.15009579/v1. Accessed 2026-09-29T16:12:30.633Z.
Source evidence and access
Original deposited abstract and posted/first-online metadata
Paraphrase of the inspected original abstract: An unreviewed study engineers glucose dehydrogenase to use the smaller NMN coenzyme, with a variant outperforming the natural enzyme–coenzyme pairing in reported activity and catalytic efficiency. Coupling the variant to four partner enzymes demonstrates NMN-recycling cascades. This broadens options for cofactor-dependent biocatalysis.
Restricted to original publisher/repository-deposited abstract and metadata in Crossref; landing page HTTP403 and full text not inspected.
Isabelle E. Petrucci; Jeanne Masson-Makdissi; Zisheng Xue; Caroline McCleary; Michal P. Glogowski; Mark D. Levin. Mechanistically Agnostic Aliphatic N–H Transmutations Enabled by Interrupted Nitrogen Deletion. Journal of the American Chemical Society; 2026; Peer-reviewed journal article, first online. DOI: 10.1021/jacs.6c13570. Accessed 2026-09-29T16:12:30.633Z.
Source evidence and access
Original deposited abstract and posted/first-online metadata
Paraphrase of the inspected original abstract: Interrupted nitrogen deletion converts heteroaryl-fused piperidines into intermediates that support three different skeletal replacements: nitrogen-to-carbonyl, nitrogen-to-oxygen and nitrogen-to-sulfur. Experiments and calculations relate productive intermediate formation to the energetic cost of disrupting aromaticity. The work expands routes between related ring systems for molecular synthesis.
Restricted to original publisher/repository-deposited abstract and metadata in Crossref; landing page HTTP403 and full text not inspected.
Publication record
Published 2026-09-29.
Sources, selection and claims were checked in an independent AI editorial review, followed by the AI editor's approval. This is not academic peer review.
- 2026-09-29 · Version e5458805 · Viewing this version