The amino acids synthesis from elementary precursors in abiotic conditions is traditionally described according to the Strecker reaction, thoroughly invoked to justify the observation of amino acids in extraterrestrial samples, and their emergence in the primordial Earth. To this day, however, a quantitative microscopic description of the mechanism, thermodynamics and kinetics of the multi-step Strecker reaction is still lacking. In the present work we tackle this study by adopting a state-of-the-art ab initio computational approach, combining an efficient scheme of exploration of the relevant chemical networks with a rigorous determination of the underlying free energy and transition states. We determine the step-by-step chemical pathway from "Strecker precursors" to glycine in solution, and calculate the corresponding full free energy landscape. Our results agree well with the scarce available experimental data, and complete them, thus providing the first end-to-end study of this complex reaction, a crucial bottleneck for the emergence of life.

The study of the thermodynamics, kinetics, and microscopic mechanisms of chemical reactions in solution requires the use of advanced free-energy methods for predictions to be quantitative. This task is however a formidable one for atomistic simulation methods, as the cost of quantum-based ab initio approaches, to obtain statistically meaningful samplings of the relevant chemical spaces and networks, becomes exceedingly heavy. In this work, we critically assess the optimal structure and minimal size of an ab initio training set able to lead to accurate free-energy profiles sampled with neural network potentials. The results allow one to propose an ab initio protocol where the ad hoc inclusion of a machine-learning (ML)-based task can significantly increase the computational efficiency, while keeping the ab initio accuracy and, at the same time, avoiding some of the notorious extrapolation risks in typical atomistic ML approaches. We focus on two representative, and computationally challenging, reaction steps of the classic Strecker-cyanohydrin mechanism for glycine synthesis in water solution, where the main precursors are formaldehyde and hydrogen cyanide. We demonstrate that indistinguishable ab initio quality results are obtained, thanks to the ML subprotocol, at about 1 order of magnitude less of computational load.

Reaction coordinates are an essential ingredient of theoretical studies of rare events in chemistry and physics because they carry information about reaction mechanism and allow the computation of free-energy landscapes and kinetic rates. We present a critical assessment of the merits and disadvantages of heuristic reaction coordinates, largely employed today, with respect to coordinates optimized on the basis of reliable transition-path sampling data. We take as a test bed multinanosecond ab initio molecular dynamics simulations of chloride SN2 substitution on methyl chloride in explicit water. The computational protocol we devise allows the unsupervised optimization of agnostic coordinates able to account for solute and solvent contributions, yielding a free-energy reconstruction of quality comparable to the best heuristic coordinates without requiring chemical intuition.