Explore the vast range of titles available.

Protein-ligand docking is an established tool in drug discovery and development to narrow down potential therapeutics for experimental testing. However, a high-quality protein structure is required and often the protein is treated as fully or partially rigid. Here we develop an AI system that can predict the fully flexible all-atom structure of protein-ligand complexes directly from sequence information. We find that classical docking methods are still superior, but depend upon having crystal structures of the target protein. In addition to predicting flexible all-atom structures, predicted confidence metrics (plDDT) can be used to select accurate predictions as well as to distinguish between strong and weak binders. The advances presented here suggest that the goal of AI-based drug discovery is one step closer, but there is still a way to go to grasp the complexity of protein-ligand interactions fully. Umol is available at: https://github.com/patrickbryant1/Umol . Here the authors report the AI system Umol that predicts flexible all-atom structures of protein-ligand complexes from sequence information, advancing AI-driven drug discovery: accurate structures and affinity can be selected from predicted confidence metrics (plDDT).

π-π stacking interactions are found throughout the proteome and have been shown to play a role in the liquid-liquid phase separation of intrinsically disordered proteins; however, the structural and energetic properties of π-π interactions that drive intra- and intermolecular protein interactions are poorly understood. In this study, we investigate the pairwise interactions of sp2-hybridized groups within proteins through an analysis of the Protein Data Bank. Along with these statistical data, small-molecule representations of these groups are simulated using molecular dynamics, while quantum mechanical and molecular mechanical calculations are used to characterize the energies of π-π interactions across their conformational distributions. Molecular dynamics is further used to simulate the folding and unfolding equilibria of small peptides enriched in sp2 groups. Ultimately, this study provides a thorough quantification of the energetics of π-stacking contacts in proteins and evaluates the strengths and limitations of different computational methods in accurately modelling these interactions.

According to the current database of natural products, over 25,000 compounds contain a vanillyl ring in their structure. The reasoning behind the high occurrence of the vanillyl ring structure seemed to be poorly understood, specifically the preference for a methoxy-substituted phenol structure as opposed to its dihydroxy analogue. To better understand this, we investigated the reaction mechanisms of two methoxyphenol structures, in syn and anti conformations, two hydroxyphenol structures, also in syn and anti conformations, and phenol as a reference structure, with acetic acid. Of the starting structures, the syn hydroxyphenol was found to be kinetically the most reactive, and formed the most stable product, while both hydroxyl-substituted phenols reacted more favorably with acetic acid than the methoxyphenols. A preference for the methoxyphenol molecule may exist as a way to hinder the formation of stable covalent bonds between natural products and cellular components. Keywords: vanillyl ring, capsaicin, gingerol, quantum chemical calculations, esterification.

Protein-ligand docking is an established tool in drug discovery and development to narrow down potential therapeutics for experimental testing. However, a high-quality protein structure is required and often the protein is treated as fully or partially rigid. Here we develop an AI system that can predict the fully flexible all-atom structure of protein-ligand complexes directly, given a multiple sequence alignment representation of the protein and a SMILES string representing the ligand. At a high accuracy threshold, unseen protein-ligand complexes can be predicted more accurately than for RoseTTAFold-AA, and at medium accuracy even classical docking methods that use known protein structures as input are surpassed. The high accuracy presented here suggests that the goal of AI-based drug discovery is one step closer, but there is still a way to go to fully grasp the complexity of protein-ligand interactions. Umol is available at: https://github.com/patrickbryant1/Umol .Competing Interest StatementThe authors have declared no competing interest.

The aim of this project is to find the most promising C6H12O molecular entities for use as a biofuel. To design such structures in a heuristic manner- gas-phase thermodynamic properties of all singlet C6H12O isomers (211 species) were calculated using G3MP2B3 ab initio composite method. For each isomerthe G3MP2B3 standard enthalpy of formation (Δf,98.15KH°calc)- higher heating value (HHV)- as well as relative molar Gibbs free energy (ΔG), standard molar entropy (S) was computed- and it was found that the computed Δf,298.15KH°calc reproduced the corresponding literature values - available only in case of 11 species - by average absolute deviation of 3.3 kJ/mol- so it can be assumed that this uncertainty can be used for the recommended Δf,298.15KH° values for the remaining molecular entities as well. These C6H12O isomers were structurally categorized into 13 subgroups according to their backbone features and functional groups. Amongst these structures- ethers have the highest HHV from which 3-ethyl tetrahydrofuran (38.3 MJ/kg) is a potential biofuel component however production from lignocellulose biomass is not yet reported and its other fuel properties has to be waited for characterization. When LD50 values were available- health risks of these compounds were also analyzed.

The most popular and universally predictive protein simulation models employ all-atom molecular dynamics (MD), but they come at extreme computational cost. The development of a universal, computationally efficient coarse-grained (CG) model with similar prediction performance has been a long-standing challenge. By combining recent deep learning methods with a large and diverse training set of all-atom protein simulations, we here develop a bottom-up CG force field with chemical transferability, which can be used for extrapolative molecular dynamics on new sequences not used during model parametrization. We demonstrate that the model successfully predicts folded structures, intermediates, metastable folded and unfolded basins, and the fluctuations of intrinsically disordered proteins while it is several orders of magnitude faster than an all-atom model. This showcases the feasibility of a universal and computationally efficient machine-learned CG model for proteins.