Prebiotic chemical reactivity in solution with quantum accuracy and microsecond sampling using neural network potentials

Proceedings of the National Academy of Sciences, Journal Year: 2024, Volume and Issue: 121(23)

Published: May 29, 2024

While RNA appears as a good candidate for the first autocatalytic systems preceding emergence of modern life, synthesis oligonucleotides without enzymes remains challenging. Because uncatalyzed reaction is extremely slow, experimental studies bring limited and indirect information on mechanism, nature which debated. Here, we develop neural network potentials (NNPs) to study phosphoester bond formation in water. NNPs are becoming routinely applied nonreactive or simple reactions, demonstrate how they can systematically be trained explore phase space complex reactions involving several proton transfers exchanges heavy atoms. We then propagate at moderate computational cost hundreds nanoseconds variety enhanced sampling simulations with quantum accuracy explicit solvent conditions. The thermodynamically preferred pathway concerted, dissociative transient metaphosphate transition state direct participation water molecules that facilitate exchange protons through nonbridging phosphate oxygens. Associative-dissociative pathways, characterized by much tighter pentacoordinated phosphate, higher free energy. Our also suggest diprotonated whose reactivity never directly assessed experiments, significantly less reactive than monoprotonated species, suggesting it probably species normal pH These observations rationalize unexplained results temperature dependence rate, pave way design more efficient abiotic catalysts activating groups.

Language: Английский

---

Enantioselective Synthesis of Pillar[5]arenes by Dynamic Kinetic Resolution via Pd-Catalyzed Suzuki Cross-Coupling

ACS Catalysis, Journal Year: 2025, Volume and Issue: unknown, P. 2607 - 2619

Published: Jan. 28, 2025

Language: Английский

---

Random Sampling Versus Active Learning Algorithms for Machine Learning Potentials of Quantum Liquid Water

Journal of Chemical Theory and Computation, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 14, 2025

Training accurate machine learning potentials requires electronic structure data comprehensively covering the configurational space of system interest. As construction this is computationally demanding, many schemes for identifying most important structures have been proposed. Here, we compare performance high-dimensional neural network (HDNNPs) quantum liquid water at ambient conditions trained to sets constructed using random sampling as well various flavors active based on query by committee. Contrary common understanding learning, find that a given set size, leads smaller test errors not included in training process. In our analysis, show can be related small energy offsets caused bias added which overcome instead correlations an error measure invariant such shifts. Still, all HDNNPs yield very similar and structural properties water, demonstrates robustness procedure with respect algorithm even when few 200 structures. However, preliminary potentials, reasonable initial avoid unnecessary extension covered configuration less relevant regions.

Language: Английский

---

On the Stability Constants of Metal-Nitrate Complexes in Aqueous Solutions

Physical Chemistry Chemical Physics, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 1, 2025

Stability constants of simple reactions involving addition the NO 3 − ion to hydrated metal complexes, [M(H 2 O) x ] n + are calculated with a computational workflow developed using cloud computing resources.

Language: Английский

---

Efficient Training of Neural Network Potentials for Chemical and Enzymatic Reactions by Continual Learning

Journal of Chemical Theory and Computation, Journal Year: 2025, Volume and Issue: unknown

Published: March 2, 2025

Machine learning (ML) methods have emerged as an efficient surrogate for high-level electronic structure theory, offering precision and computational efficiency. However, the vast conformational chemical space remains challenging when constructing a general force field. Training data sets typically cover only limited region of this space, resulting in poor extrapolation performance. Traditional strategies must address problem by training models from scratch using old new sets. In addition, model transferability is crucial field construction. Existing ML fields, designed closed systems with no external environmental potential, exhibit to complex condensed phase such enzymatic reactions, inferior performance high memory costs. Our ML/MM model, based on Taylor expansion electrostatic operator, showed between reactions several simple solvents. This work extends strategy explore more heterogeneous environments. we also apply continual enable autonomous on-the-fly continuous stream data. By combining these two methods, can efficiently construct that be applied various media.

Language: Английский

---

: A Toolkit for Autonomous, User-Guided Construction of Machine-Learned Potential Energy Surfaces

Computer Physics Communications, Journal Year: 2024, Volume and Issue: unknown, P. 109446 - 109446

Published: Nov. 1, 2024

Language: Английский

---

Data efficient machine learning potentials for modeling catalytic reactivity via active learning and enhanced sampling

npj Computational Materials, Journal Year: 2024, Volume and Issue: 10(1)

Published: Dec. 19, 2024

Language: Английский

---