MoleGuLAR: Molecule Generation Using Reinforcement Learning with Alternating Rewards.

The design of new inhibitors for novel targets is a very important problem especially in the current scenario with the world being plagued by COVID-19. Conventional approaches such as high-throughput virtual screening require extensive combing through existing data sets in the hope of finding possible matches. In this study, we propose a computational strategy for de novo generation of molecules with high binding affinities to the specified target and other desirable properties for druglike molecules using reinforcement learning. A deep generative model built using a stack-augmented recurrent neural network initially trained to generate druglike molecules is optimized using reinforcement learning to start generating molecules with desirable properties like LogP, quantitative estimate of drug likeliness, topological polar surface area, and hydration free energy along with the binding affinity. For multiobjective optimization, we have devised a novel strategy in which the property being used to calculate the reward is changed periodically. In comparison to the conventional approach of taking a weighted sum of all rewards, this strategy shows an enhanced ability to generate a significantly higher number of molecules with desirable properties.

Multiscale Modeling of Wobble to Watson-Crick-Like Guanine-Uracil Tautomerization Pathways in RNA.

Energetically unfavorable Watson-Crick (WC)-like tautomeric forms of nucleobases are known to introduce spontaneous mutations, and contribute to replication, transcription, and translation errors. Recent NMR relaxation dispersion techniques were able to show that wobble (w) G•U mispair exists in equilibrium with the short-lived, low-population WC-like enolic tautomers. Presently, we have investigated the wG•U → WC-like enolic reaction pathway using various theoretical methods: quantum mechanics (QM), molecular dynamics (MD), and combined quantum mechanics/molecular mechanics (QM/MM). The previous studies on QM gas phase calculations were inconsistent with experimental data. We have also explored the environmental effects on the reaction energies by adding explicit water. While the QM-profile clearly becomes endoergic in the presence of water, the QM/MM-profile remains consistently endoergic in the presence and absence of water. Hence, by including microsolvation and QM/MM calculations, the experimental data can be explained. For the G•Uenol→ Genol•U pathway, the latter appears to be energetically more favorable throughout all computational models. This study can be considered as a benchmark of various computational models of wG•U to WC-like tautomerization pathways with and without the environmental effects, and may contribute on further studies of other mispairs as well.

Desolvation of Peptide Bond by O to S Substitution Impacts Protein Stability.

Amino acid side chains are key to fine-tuning the microenvironment polarity in proteins composed of polar amide bonds. Here, we report that substituting an oxygen atom of the backbone amide bond with sulfur atom desolvates the thioamide bond, thereby increasing its lipophilicity. The impact of such local desolvation by O to S substitution in proteins was tested by synthesizing thioamidated variants of Pin1 WW domain. We observe that a thioamide acts in synergy with nonpolar amino acid side chains to reduce the microenvironment polarity and increase protein stability by more than 14 °C. Through favorable van der Waals and hydrogen bonding interactions, this single atom substitution significantly stabilizes proteins without altering the amino acid sequence and structure of the native protein.

Transition between [R]- and [S]-stereoisomers without bond breaking.

The fifty-year old proposal of a nondissociative racemization reaction of a tetracoordinated tetrahedral center from one enantiomer to another via a planar transition state by Hoffmann and coworkers has been explored by many research groups over the past five decades. A number of stable molecules with planar tetracoordinated and higher-coordinated centers have been designed and experimentally realized; however, there has not been a single example of a molecular system that can possibly undergo such racemization. Here we show examples of molecular species that undergo inversion of stereochemistry around tetrahedral centers (Si, Al- and P+) either via a planar transition state or an intermediate state using quantum mechanical, ab initio quasi-classical dynamics calculations, and Born-Oppenheimer molecular dynamics (BOMD) simulations. This work is expected to provide potential leads for future studies on this fundamental phenomenon in chemistry.

Machine Learning for Accurate Force Calculations in Molecular Dynamics Simulations.

The computationally expensive nature of ab initio molecular dynamics simulations severely limits its ability to simulate large system sizes and long time scales, both of which are necessary to imitate experimental conditions. In this work, we explore an approach to make use of the data obtained using the quantum mechanical density functional theory (DFT) on small systems and use deep learning to subsequently simulate large systems by taking liquid argon as a test case. A suitable vector representation was chosen to represent the surrounding environment of each Ar atom, and a Δ-NetFF machine learning model, where the neural network was trained to predict the difference in resultant forces obtained by DFT and classical force fields, was introduced. Molecular dynamics simulations were then performed using forces from the neural network for various system sizes and time scales depending on the properties we calculated. A comparison of properties obtained from the classical force field and the neural network model was presented alongside available experimental data to validate the proposed method.

Limits of the creation of electronic wave packets using time-dependent density functional theory.

Explicitly time-dependent density functional theory (TDDFT) has often been suggested as the method of choice for controlling the correlated dynamics of many electron systems. However, it is not yet clear which control tasks can be achieved reliably and how this depends on the functionals used. In this article, we show that the control task of creating a simple wave packet, having a population of 50% in the excited state, can indeed be achieved if a certain condition is fulfilled. This result is in contrast to the observation that a full population inversion is extremely difficult to achieve. In addition, we identify a rule to predict when TDDFT produces the correct wave packet. To illustrate our findings, we study the molecules Li(2)C(2), Li(7)OH, and B(2)N(2)CO using two different functionals as well as time-dependent Hartree-Fock (TDHF). To assess the performance of TDDFT and TDHF, we compare with time-dependent configuration interaction calculations.

Coherent control and time-dependent density functional theory: towards creation of wave packets by ultrashort laser pulses.

Explicitly time-dependent density functional theory (TDDFT) is a formally exact theory, which can treat very large systems. However, in practice it is used almost exclusively in the adiabatic approximation and with standard ground state functionals. Therefore, if combined with coherent control theory, it is not clear which control tasks can be achieved reliably, and how this depends on the functionals. In this paper, we continue earlier work in order to establish rules that answer these questions. Specifically, we look at the creation of wave packets by ultrashort laser pulses that contain several excited states. We find that (i) adiabatic TDDFT only works if the system is not driven too far from the ground state, (ii) the permanent dipole moments involved should not differ too much, and (iii) these results are independent of the functional used. Additionally, we find an artifact that produces fluence-dependent excitation energies.

SCONES: Self-Consistent Neural Network for Protein Stability Prediction Upon Mutation.

Engineering proteins to have desired properties by mutating amino acids at specific sites is commonplace. Such engineered proteins must be stable to function. Experimental methods used to determine stability at throughputs required to scan the protein sequence space thoroughly are laborious. To this end, many machine learning based methods have been developed to predict thermodynamic stability changes upon mutation. These methods have been evaluated for symmetric consistency by testing with hypothetical reverse mutations. In this work, we propose transitive data augmentation, evaluating transitive consistency with our new Stransitive data set, and a new machine learning based method, the first of its kind, that incorporates both symmetric and transitive properties into the architecture. Our method, called SCONES, is an interpretable neural network that predicts small relative protein stability changes for missense mutations that do not significantly alter the structure. It estimates a residue's contributions toward protein stability (ΔG) in its local structural environment, and the difference between independently predicted contributions of the reference and mutant residues is reported as ΔΔG. We show that this self-consistent machine learning architecture is immune to many common biases in data sets, relies less on data than existing methods, is robust to overfitting, and can explain a substantial portion of the variance in experimental data.

Polarizable Multipolar Molecular Dynamics Using Distributed Point Charges.

Distributed point charge models (DCM) and their minimal variants (MDCM) have been integrated with tools widely used for condensed-phase simulations, including a virial-based barostat and a slow-growth algorithm for thermodynamic integration. Minimal DCM is further developed in a systematic fashion to reduce fitting errors in the electrostatic interaction energy, and a new fragment-based approach offers considerable speedup of the MDCM fitting process for larger molecules with increased numbers of off-centered charged sites. Finally, polarizable (M)DCM is also introduced in the present work. The developments are used in condensed-phase simulations of popular force fields with commonly applied simulation conditions. (M)DCM equivalents for a range of widely used water force fields and for fluorobenzene (PhF) are developed and applied along with the original models to evaluate the impact of reformulating the electrostatic term. Comparisons of the molecular electrostatic potential (MEP), electrostatic interaction energies, and bulk properties from molecular dynamics simulations for a range of models from simple TIPnP (n = 3-5) to the polarizable, multipolar iAMOEBA models for water and an existing quadrupolar model for PhF confirm that DCMs retain the accuracy of the original models, providing a homogeneous, efficient, and generic point charge alternative to a multipolar electrostatic model for force field development and multilevel simulations.

Urea-aromatic interactions in biology.

Noncovalent interactions are key determinants in both chemical and biological processes. Among such processes, the hydrophobic interactions play an eminent role in folding of proteins, nucleic acids, formation of membranes, protein-ligand recognition, etc.. Though this interaction is mediated through the aqueous solvent, the stability of the above biomolecules can be highly sensitive to any small external perturbations, such as temperature, pressure, pH, or even cosolvent additives, like, urea-a highly soluble small organic molecule utilized by various living organisms to regulate osmotic pressure. A plethora of detailed studies exist covering both experimental and theoretical regimes, to understand how urea modulates the stability of biological macromolecules. While experimentalists have been primarily focusing on the thermodynamic and kinetic aspects, theoretical modeling predominantly involves mechanistic information at the molecular level, calculating atomistic details applying the force field approach to the high level electronic details using the quantum mechanical methods. The review focuses mainly on examples with biological relevance, such as (1) urea-assisted protein unfolding, (2) urea-assisted RNA unfolding, (3) urea lesion interaction within damaged DNA, (4) urea conduction through membrane proteins, and (5) protein-ligand interactions those explicitly address the vitality of hydrophobic interactions involving exclusively the urea-aromatic moiety.

The Role of Water in the Stability of Wild-type and Mutant Insulin Dimers.

Insulin dimerization and aggregation play important roles in the endogenous delivery of the hormone. One of the important residues at the insulin dimer interface is PheB24, which is an invariant aromatic anchor that packs toward its own monomer inside a hydrophobic cavity formed by ValB12, LeuB15, TyrB16, CysB19, and TyrB26. Using molecular dynamics and free-energy simulations within explicit solvent, the structural and dynamical consequences of mutations of Phe at position B24 to glycine (Gly), alanine (Ala), and d-Ala and the des-PheB25 variant are quantified. Consistent with experiments, it is found that the Gly and Ala modifications lead to insulin dimers with reduced stability by 4 and 5 kcal/mol from thermodynamic integration and 4 and 8 kcal/mol from results using molecular mechanics-generalized Born surface area, respectively. Given the experimental difficulties to quantify the thermodynamic stability of modified insulin dimers, such computations provide a valuable complement. Interestingly, the Gly mutant exists as a strongly and a weakly interacting dimer. Analysis of the molecular dynamics simulations shows that this can be explained by water molecules that replace direct monomer-monomer H-bonding contacts at the dimerization interface involving residues B24 to B26. It is concluded that such solvent molecules play an essential role and must be included in future insulin dimerization studies.

A Novel, Computationally Efficient Multipolar Model Employing Distributed Charges for Molecular Dynamics Simulations.

A truncated multipole expansion can be re-expressed exactly using an appropriate arrangement of point charges. This means that groups of point charges that are shifted away from nuclear coordinates can be used to achieve accurate electrostatics for molecular systems. We introduce a multipolar electrostatic model formulated in this way for use in computationally efficient multipolar molecular dynamics simulations with well-defined forces and energy conservation in NVE (constant number-volume-energy) simulations. A framework is introduced to distribute torques arising from multipole moments throughout a molecule, and a refined fitting approach is suggested to obtain atomic multipole moments that are optimized for accuracy and numerical stability in a force field context. The formulation of the charge model is outlined as it has been implemented into CHARMM, with application to test systems involving H2O and chlorobenzene. As well as ease of implementation and computational efficiency, the approach can be used to provide snapshots for multipolar QM/MM calculations in QM/MM-MD studies and easily combined with a standard point-charge force field to allow mixed multipolar/point charge simulations of large systems.

The Lack of Resonance Problem in Coherent Control with Real-Time Time-Dependent Density Functional Theory.

We will show that adiabatic real-time TDDFT predicts a time- and energy-dependent electronic structure of molecules, which makes it hard to combine TDDFT with coherent control schemes that depend on resonance conditions. In this study, we use sequences of ultrashort pulses, separated by long intervals of field free evolution, to illustrate this phenomenon for two molecules and two functionals. In coherent control scenarios, long laser pulses, with many oscillation periods, excite the system continuously and, in this way produce, a time-dependent electronic structure.

Critical Examination of Explicitly Time-Dependent Density Functional Theory for Coherent Control of Dipole Switching.

We compare the performance of explicitly time-dependent density functional theory (DFT) with time-dependent configuration interaction (TDCI) to achieve the control task of a population inversion in LiCN. We assume that if a given pulse achieves the control task when used in TDCI, then there should be a pulse with similar frequency and intensity that achieves the task in time-dependent DFT (TDDFT). The present investigation indicates that this is not the case, if standard functionals are used in the adiabatic approximation.