A Theoretical Study on the Structural, Electronic, and Magnetic Properties of Bimetallic Pt13-nNin (N = 0, 3, 6, 9, 13) Nanoclusters to Unveil the Catalytic Mechanisms for the Water-Gas Shift Reaction

In this work, first-principles calculations by using density functional theory at the GFN-xTB level, are performed to investigate the relative stability and structural, electronic, and magnetic properties of bimetallic Pt13-nNin (n = 0, 3, 6, 9, 13) nanoclusters by using corrected Hammer and Nørskov model. In addition, by employing the reaction path and the energetic span models, the energy profile and the turnover frequency are calculated to disclose the corresponding reaction mechanism of the water-gas shift reaction catalyzed by these nanoclusters. Our findings render that Ni causes an overall shrinking of the nanocluster's size and misalignment of the spin channels, increasing the magnetic nature of the nanoclusters. Pt7Ni6 nanocluster is the most stable as a result of the better coupling between the Pt and Ni d-states. Pt4Ni9 maintains its structure over the reaction cycle, with a larger turnover frequency value than Pt7Ni6. On the other hand, despite Pt10Ni3 presenting the highest value of turnover frequency, it suffers a strong structural deformation over the completion of a reaction cycle, indicating that the catalytic activity can be altered.

A Neural Network Approach for Property Determination of Molecular Solar Cell Candidates

The dihydroazulene/vinylheptafulvene (DHA/VHF) photocouple is a promising candidate for molecular solar heat batteries, storing and releasing energy in a closed cycle. Much work has been done on improving the energy storage capacity and the half-life of the high-energy isomer via substituent functionalization, but similarly important is keeping these improved properties in common polar solvents, along with being soluble in these, which is tied to the dipole properties. However, the number of possible derivatives makes an overview of this combinatorial space impossible both for experimental work and traditional computational chemistry. Due to the time-consuming nature of running many thousands of computations, we look to machine learning, which bears the advantage that once a model has been trained, it can be used to rapidly estimate approximate values for the given system. Applying a convolutional neural network, we show that it is possible to reach good agreement with traditional computations on a scale that allows us to rapidly screen tens of thousands of the DHA/VHF photocouple, eliminating bad candidates and allowing computational resources to be directed toward meaningful compounds.

Fast and automated identification of reactions with low barriers using meta-MD simulations

We test our meta-molecular dynamics (MD) based approach for finding low-barrier (<30 kcal/mol) reactions on uni- and bimolecular reactions extracted from the barrier dataset developed by Grambow, Pattanaik & Green (2020). For unimolecular reactions the meta-MD simulations identify 25 of the 26 products found by Grambow, Pattanaik & Green (2020), while the subsequent semiempirical screening eliminates an additional four reactions due to an overestimation of the reaction energies or estimated barrier heights relative to DFT. In addition, our approach identifies 36 reactions not found by Grambow, Pattanaik & Green (2020), 10 of which are <30 kcal/mol. For bimolecular reactions the meta-MD simulations identify 19 of the 20 reactions found by Grambow, Pattanaik & Green (2020), while the subsequent semiempirical screening eliminates an additional reaction. In addition, we find 34 new low-barrier reactions. For bimolecular reactions we found that it is necessary to “encourage” the reactants to go to previously undiscovered products, by including products found by other MD simulations when computing the biasing potential as well as decreasing the size of the molecular cavity in which the MD occurs, until a reaction is observed. We also show that our methodology can find the correct products for two reactions that are more representative of those encountered in synthetic organic chemistry. The meta-MD hyperparameters used in this study thus appear to be generally applicable to finding low-barrier reactions.

Generating stable molecules using imitation and reinforcement learning

Chemical space is routinely explored by machine learning methods to discover interesting molecules, before time-consuming experimental synthesizing is attempted. However, these methods often rely on a graph representation, ignoring 3D information necessary for determining the stability of the molecules. We propose a reinforcement learning (RL) approach for generating molecules in Cartesian coordinates allowing for quantum chemical prediction of the stability. To improve sample-efficiency we learn basic chemical rules from imitation learning (IL) on the GDB-11 database to create an initial model applicable for all stoichiometries. We then deploy multiple copies of the model conditioned on a specific stoichiometry in a RL setting. The models correctly identify low energy molecules in the database and produce novel isomers not found in the training set. Finally, we apply the model to larger molecules to show how RL further refines the IL model in domains far from the training data.

Grammatical evolution-based design of SARS-CoV-2 main protease inhibitors

A series of SARS-CoV-2 main protease (SARS-CoV-2-M^pro) inhibitors were modeled using evolutive grammar algorithms. We have generated an automated program that finds the best candidate to inhibit the main protease, M^pro, of SARS-CoV-2. The candidates were constructed based on a pharmacophore model of the above-mentioned target; relevant moieties of such molecules were modified using data-basis sets with similar chemical behavior to the reference moieties. Additionally, we used the SMILES language to translate 3D chemical structures to 1D words; then, an evolutive grammar algorithm was used to explore the chemical space and obtain new candidates, which were evaluated via the binding energy of molecular coupling assays as an evaluation function. Finally, sixteen molecules were obtained in 3 runs of our program, three of which show promising binding properties as SARS-CoV-2-M^pro inhibitors. One of them, TTO, maintained its relevant binding properties during 100 ns molecular dynamics experiments. For this reason, TTO is the best candidate to inhibit SARS-CoV-2-M^pro. The software we developed for this contribution is available at the following URL: https://github.com/masotelof/GEMolecularDesign.

LEADD: Lamarckian evolutionary algorithm for de novo drug design

Given an objective function that predicts key properties of a molecule, goal-directed de novo molecular design is a useful tool to identify molecules that maximize or minimize said objective function. Nonetheless, a common drawback of these methods is that they tend to design synthetically unfeasible molecules. In this paper we describe a Lamarckian evolutionary algorithm for de novo drug design (LEADD). LEADD attempts to strike a balance between optimization power, synthetic accessibility of designed molecules and computational efficiency. To increase the likelihood of designing synthetically accessible molecules, LEADD represents molecules as graphs of molecular fragments, and limits the bonds that can be formed between them through knowledge-based pairwise atom type compatibility rules. A reference library of drug-like molecules is used to extract fragments, fragment preferences and compatibility rules. A novel set of genetic operators that enforce these rules in a computationally efficient manner is presented. To sample chemical space more efficiently we also explore a Lamarckian evolutionary mechanism that adapts the reproductive behavior of molecules. LEADD has been compared to both standard virtual screening and a comparable evolutionary algorithm using a standardized benchmark suite and was shown to be able to identify fitter molecules more efficiently. Moreover, the designed molecules are predicted to be easier to synthesize than those designed by other evolutionary algorithms.

Cu(II)(PhOMe-Salophen) Complex: Greener Pasture Biological Study, XRD/HAS Interactions, and MEP

PhOMe-salophen (1b) (salophen is N,N-bis(salycilidene)-1,2-phenylenediamine with two tert-butyl on each ring) and Cu(II) complex with PhOMe-salophen (1c) have been synthesized and characterized using various tools, including X-ray diffraction for the Cu(II)-complex (1c, C₄₃H₅₂CuN₂O₃)). The copper complex has been obtained by Cu²⁺ templated approach using 1b. PhOMe-salophen (1b) has been obtained in reasonably high yield using a mixture of the Schiff-base, 1a, Pd(OAc)₂, PPh₃, Na₂CO₃, 4-methoxyphenylboronic acid in benzene. We focus in this research work on the electronic and structural properties of the Cu–Schiff base complex. The tetra-coordinate τ₄ index was calculated, indicating almost a perfect square planner in agreement with X-ray diffraction results. MEP reveals the maximum positive regions in 1/-associated with the azomethine and methoxyphenyl C–H bonds with an average value of 0.03 a.u. Hirshfeld surface analysis (HSA) was also studied to highlight the significant inter-atomic contacts and their percentage contribution through 2D Fingerprint plot. In a fair comparative molecular docking study, 1b and 1c were docked together with N-[{(5-methylisoxazol-3-yl)-carbonyl}alanyl}-l-valyl]-N1-((1R,2Z)-4-(benzyloxy)-4-oxo-1-[{(3R)-2-oxopyrrolidin-3-yl}methyl]but-2-enyl)-l-leucinamide, N3 against main protease M^pro, (PDB code 7BQY) using the same parameters and conditions. Interesting here to use the free energy, in silico, molecular docking approach, which aims to rank our molecules with respect to the well-known inhibitor, N3. The binding scores of 1b, 1c, N3 are –7.8, –9.0, and –8.4 kcal/mol, respectively. These preliminary results propose that ligands deserve additional study in the context of possible remedial agents for COVID-19.

Using deep neural networks to explore chemical space

INTRODUCTION

The popularity and success of advanced AI methods like deep neural networks has led to novel ways for exploring chemical space. Their opaque nature poses challenges for model evaluation regarding novelty, uniqueness, and distribution of the chemical space covered. However, these methods also promise to be able to explore uncharted chemical space in novel ways that do not rely directly on structural similarity.

AREAS COVERED

This review provides an overview of popular deep learning methods for chemical space exploration. Crucial aspects like choice of molecular representation, training for focused chemical space exploration, and criteria for assessing and validating chemical space coverage are discussed.

EXPERT OPINION

Deep learning offers great potential for chemical space exploration beyond conventional fragment-based methods. Given the rarity of prospective applications and considering the difficulty in assessing representativeness and comprehensiveness of chemical space covered, developing criteria for assessing and validating generative models is of great significance. Latent space models like variational autoencoders are conceptually appealing for inverse QSAR/QSPR approaches as neighborhood relationships in latent space can be trained to reflect property similarities. Future research in understanding and interpreting generative models might lead to a better understanding of biologically relevant properties of molecules.

Chemical reactivity prediction: current methods and different application areas

The ability to predict chemical reactivity of a molecule is highly desirable in drug discovery, both ex vivo (synthetic route planning, formulation, stability) and in vivo: metabolic reactions determine pharmacodynamics, pharmacokinetics and potential toxic effects, and early assessment of liabilities is vital to reduce attrition rates in later stages of development. Quantum mechanics offer a precise description of the interactions between electrons and orbitals in the breaking and forming of new bonds. Modern algorithms and faster computers have allowed the study of more complex systems in a punctual and accurate fashion, and answers for chemical questions around stability and reactivity can now be provided. Through machine learning, predictive models can be built out of descriptors derived from quantum mechanics and cheminformatics, even in the absence of experimental data to train on. In this article, current progress on computational reactivity prediction is reviewed: applications to problems in drug design, such as modelling of metabolism and covalent inhibition, are highlighted and unmet challenges are posed.

Data-driven modeling of S0 → S1 excitation energy in the BODIPY chemical space: High-throughput computation, quantum machine learning, and inverse design

Derivatives of BODIPY are popular fluorophores due to their synthetic feasibility, structural rigidity, high quantum yield, and tunable spectroscopic properties. While the characteristic absorption maximum of BODIPY is at 2.5 eV, combinations of functional groups and substitution sites can shift the peak position by ±1 eV. Time-dependent long-range corrected hybrid density functional methods can model the lowest excitation energies offering a semi-quantitative precision of ±0.3 eV. Alas, the chemical space of BODIPYs stemming from combinatorial introduction of-even a few dozen-substituents is too large for brute-force high-throughput modeling. To navigate this vast space, we select 77 412 molecules and train a kernel-based quantum machine learning model providing <2% hold-out error. Further reuse of the results presented here to navigate the entire BODIPY universe comprising over 253 giga (253 × 10⁹) molecules is demonstrated by inverse-designing candidates with desired target excitation energies.

Perspective and challenges in electrochemical approaches for reactive CO2 separations

The desire toward decarbonization and renewable energy has sparked research interests in reactive CO2 separations, such as direct air capture that utilize electricity as opposed to conventional thermal and pressure swing processes, which are energy-intensive, cost-prohibitive, and fossil-fuel dependent. Although the electrochemical approaches in CO2 capture that support negative emissions technologies are promising in terms of modularity, smaller footprint, mild reaction conditions, and possibility to integrate into conversion processes, their practice depends on the wider availability of renewable electricity. This perspective discusses key advances made in electrolytes and electrodes with redox-active moieties that reversibly capture CO2 or facilitate its transport from a CO2-rich side to a CO2-lean side within the last decade. In support of the discovery of new heterogeneous electrode materials and electrolytes with redox carriers, the role of computational chemistry is also discussed.

Virtual screening of norbornadiene-based molecular solar thermal energy storage systems using a genetic algorithm

We present a computational methodology for the screening of a chemical space of 10²⁵ substituted norbornadiene molecules for promising kinetically stable molecular solar thermal (MOST) energy storage systems with high energy densities that absorb in the visible part of the solar spectrum. We use semiempirical tight-binding methods to construct a dataset of nearly 34 000 molecules and train graph convolutional networks to predict energy densities, kinetic stability, and absorption spectra and then use the models together with a genetic algorithm to search the chemical space for promising MOST energy storage systems. We identify 15 kinetically stable molecules, five of which have energy densities greater than 0.45 MJ/kg, and the main conclusion of this study is that the largest energy density that can be obtained for a single norbornadiene moiety with the substituents considered here, while maintaining a long half-life and absorption in the visible spectrum, is around 0.55 MJ/kg.

Remote Learning of COVID-19 Kinetic Analysis in a Physical Chemistry Laboratory Class

The COVID-19 pandemic has affected many in-person laboratory courses across the world. The viral spreading model is complicated but parameters, such as its reproduction number, R _t, can be estimated with the susceptible, infectious, or recovered model. COVID-19 data for many states and countries are widely available online. This provides an opportunity for the students to analyze its spreading kinetics remotely. Here, we reported a laboratory set up online during the third week of the spring semester of 2021 to minimize social contacts. Due to the wide interest in developing online physical chemistry and analytical laboratories during the pandemic, we would like to share this laboratory design. The method, technique, procedure, and grading are described in this report. The student participants were able to apply the kinetic techniques learned in physical chemistry to successfully analyze an ongoing real-world problem through a remote learning environment and prepare this report.

Explicit formulae for the peak time of an epidemic from the SIR model. Which approximant to use?

An analytic evaluation of the peak time of a disease allows for the installment of effective epidemic precautions. Recently, an explicit analytic, approximate expression (MT) for the peak time of the fraction of infected persons during an outbreak within the susceptible-infectious-recovered/removed (SIR) model had been presented and discussed (Turkyilmazoglu, 2021). There are three existing approximate solutions (SK-I, SK-II, and CG) of the semi-time SIR model in its reduced formulation that allow one to come up with different explicit expressions for the peak time of the infected compartment (Schlickeiser and Kröger, 2021; Carvalho and Gonçalves, 2021). Here we compare the four expressions for any choice of SIR model parameters and find that SK-I, SK-II and CG are more accurate than MT as long as the amount of population to which the SIR model is applied exceeds hundred by far (countries, ss, cities). For small populations with less than hundreds of individuals (families, small towns), however, the approximant MT outperforms the other approximants. To be able to compare the various approaches, we clarify the equivalence between the four-parametric dimensional SIR equations and their two-dimensional dimensionless analogue. Using Covid-19 data from various countries and sources we identify the relevant regime within the parameter space of the SIR model.

MOLGENGO: Finding Novel Molecules with Desired Electronic Properties by Capitalizing on Their Global Optimization

The discovery of novel and favorable fluorophores is critical for understanding many chemical and biological studies. High-resolution biological imaging necessitates fluorophores with diverse colors and high quantum yields. The maximum oscillator strength and its corresponding absorption wavelength of a molecule are closely related to the quantum yields and the emission spectrum of fluorophores, respectively. Thus, the core step to design favorable fluorophore molecules is to optimize the desired electronic transition properties of molecules. Here, we present MOLGENGO, a new molecular property optimization algorithm, to discover novel and favorable fluorophores with machine learning and global optimization. This study reports novel molecules from MOLGENGO with high oscillator strength and absorption wavelength close to 200, 400, and 600 nm. The results of MOLGENGO simulations have the potential to be candidates for new fluorophore frameworks.

Chemistrees: Data-Driven Identification of Reaction Pathways via Machine Learning

We propose to analyze molecular dynamics (MD) output via a supervised machine learning (ML) algorithm, the decision tree. The approach aims to identify the predominant geometric features which correlate with trajectories that transition between two arbitrarily defined states. The data-driven algorithm aims to identify these features without the bias of human “chemical intuition”. We demonstrate the method by analyzing the proton exchange reactions in formic acid solvated in small water clusters. The simulations were performed with ab initio MD combined with a method to efficiently sample the rare event, path sampling. Our ML analysis identified relevant geometric variables involved in the proton transfer reaction and how they may change as the number of solvating water molecules changes.

Targeted Free Energy Perturbation Revisited: Accurate Free Energies from Mapped Reference Potentials

We present an approach that extends the theory of targeted free energy perturbation (TFEP) to calculate free energy differences and free energy surfaces at an accurate quantum mechanical level of theory from a cheaper reference potential. The convergence is accelerated by a mapping function that increases the overlap between the target and the reference distributions. Building on recent work, we show that this map can be learned with a normalizing flow neural network, without requiring simulations with the expensive target potential but only a small number of single-point calculations, and, crucially, avoiding the systematic error that was found previously. We validate the method by numerically evaluating the free energy difference in a system with a double-well potential and by describing the free energy landscape of a simple chemical reaction in the gas phase.

Mechanistic and thermodynamic characterization of oxathiazolones as potent and selective covalent immunoproteasome inhibitors

The ubiquitin–proteasome system is responsible for the degradation of proteins and plays a critical role in key cellular processes. While the constitutive proteasome (cPS) is expressed in all eukaryotic cells, the immunoproteasome (iPS) is primarily induced during disease processes, and its inhibition is beneficial in the treatment of cancer, autoimmune disorders and neurodegenerative diseases. Oxathiazolones were reported to selectively inhibit iPS over cPS, and the inhibitory activity of several oxathiazolones against iPS was experimentally determined. However, the detailed mechanism of the chemical reaction leading to irreversible iPS inhibition and the key selectivity drivers are unknown, and separate characterization of the noncovalent and covalent inhibition steps is not available for several compounds. Here, we investigate the chemical reaction between oxathiazolones and the Thr1 residue of iPS by quantum mechanics/molecular mechanics (QM/MM) simulations to establish a plausible reaction mechanism and to determine the rate-determining step of covalent complex formation. The modelled binding mode and reaction mechanism are in line with the selective inhibition of iPS versus cPS by oxathiazolones. The k_inact value of several ligands was estimated by constructing the potential of mean force of the rate-determining step by QM/MM simulations coupled with umbrella sampling. The equilibrium constant K_i of the noncovalent complex formation was evaluated by classical force field-based thermodynamic integration. The calculated K_i and k_inact values made it possible to analyse the contribution of the noncovalent and covalent steps to the overall inhibitory activity. Compounds with similar intrinsic reactivities exhibit varying selectivities for iPS versus cPS owing to subtle differences in the binding modes that slightly affect K_i, the noncovalent affinity, and importantly alter k_inact, the covalent reactivity of the bound compounds. A detailed understanding of the inhibitory mechanism of oxathiazolones is useful in designing iPS selective inhibitors with improved drug-like properties.

The spreading of Covid-19 in Mexico: A diffusional approach

In this work, we analyze the spreading of Covid-19 in Mexico using the spatial SEIRD epidemiologic model. We use the information of the 32 regions (States) that conform the country, such as population density, verified infected cases, and deaths in each State. We extend the SEIRD compartmental epidemiologic with diffusion mechanisms in the exposed and susceptible populations. We use the Fickian law with the diffusion coefficient proportional to the population density to encompass the diffusion effects. The numerical results suggest that the epidemiologic model demands time-dependent parameters to incorporate non-monotonous behavior in the actual data in the global dynamic. The diffusional model proposed in this work has great potential in predicting the virus spreading on different scales, i.e., local, national, and between countries, since the complete reduction in people mobility is impossible.

Explicit formulae for the peak time of an epidemic from the SIR model

Reducing the peak time of an epidemic disease in order for slowing down the eventual dynamics and getting prepared for the unavoidable epidemic wave is utmost significant to fight against the risks of a contagious epidemic disease. To serve to this purpose, the well-documented infection model of SIR is examined in the current research to propose an analytical approach for providing an explicit formula associated with a straightforward computation of peak time of outbreak. Initially, the time scale from the relevant autonomous SIR epidemic model is formulated analytically via an integral based on the fractions of susceptible and infected compartments. Afterwards, through a series expansion of the logarithmic term of the resultant integrand, the peak time is shown to rely upon the fraction of susceptible, the infectious ratio as well as the initial fractions of ill and susceptible individuals. The approximate expression is shown to rigorously capable of capturing the time threshold of illness for an epidemic from the semi-time SIR epidemiology. Otherwise, it is also successful to predict the peak time from a past history of a disease when all-time epidemic model is adopted. Accuracy of the derived expressions are initially confirmed by direct comparisons with recently reported approximate formulas in the literature. Several other epidemic disease samples including the COVID-19 often studied in the recent literature are eventually attacked with favourable performance of the presented formulae for assessing the peak time occurrence of an epidemic. A quick evaluation of the peak time of a disease certainly enables the governments to take early effective epidemic precautions.

The impact of curcumin derived polyphenols on the structure and flexibility COVID-19 main protease binding pocket: a molecular dynamics simulation study

The newly occurred SARS-CoV-2 caused a leading pandemic of coronavirus disease (COVID-19). Up to now it has infected more than one hundred sixty million and killed more than three million people according to 14 May 2021 World Health Organization report. So far, different types of studies have been conducted to develop an anti-viral drug for COVID-19 with no success yet. As part of this, silico were studied to discover and introduce COVID-19 antiviral drugs and results showed that protease inhibitors could be very effective in controlling. This study aims to investigate the binding affinity of three curcumin derived polyphenols against COVID-19 the main protease (Mpro), binding pocket, and identification of important residues for interaction. In this study, molecular modeling, auto-dock coupled with molecular dynamics simulations were performed to analyze the conformational, and stability of COVID-19 binding pocket with diferuloylmethane, demethoxycurcumin, and bisdemethoxycurcumin. All three compounds have shown binding affinity −39, −89 and −169.7, respectively. Demethoxycurcumin and bisdemethoxycurcumin showed an optimum binding affinity with target molecule and these could be one of potential ligands for COVID-19 therapy. And also, COVID-19 main protease binding pocket binds with the interface region by one hydrogen bond. Moreover, the MD simulation parameters indicated that demethoxycurcumin and bisdemethoxycurcumin were stable during the simulation run. These findings can be used as a baseline to develop therapeutics with curcumin derived polyphenols against COVID-19.

Deep molecular dreaming: inverse machine learning for de-novo molecular design and interpretability with surjective representations

Computer-based de-novo design of functional molecules is one of the most prominent challenges in cheminformatics today. As a result, generative and evolutionary inverse designs from the field of artificial intelligence have emerged at a rapid pace, with aims to optimize molecules for a particular chemical property. These models ‘indirectly’ explore the chemical space; by learning latent spaces, policies, and distributions, or by applying mutations on populations of molecules. However, the recent development of the SELFIES (Krenn 2020 Mach. Learn.: Sci. Technol. 1 045024) string representation of molecules, a surjective alternative to SMILES, have made possible other potential techniques. Based on SELFIES, we therefore propose PASITHEA, a direct gradient-based molecule optimization that applies inceptionism (Mordvintsev 2015) techniques from computer vision. PASITHEA exploits the use of gradients by directly reversing the learning process of a neural network, which is trained to predict real-valued chemical properties. Effectively, this forms an inverse regression model, which is capable of generating molecular variants optimized for a certain property. Although our results are preliminary, we observe a shift in distribution of a chosen property during inverse-training, a clear indication of PASITHEA’s viability. A striking property of inceptionism is that we can directly probe the model’s understanding of the chemical space on which it is trained. We expect that extending PASITHEA to larger datasets, molecules and more complex properties will lead to advances in the design of new functional molecules as well as the interpretation and explanation of machine learning models.

Search for Non-Protein Protease Inhibitors Constituted with an Indole and Acetylene Core

A possible inhibitor of proteases, which contains an indole core and an aromatic polar acetylene, was designed and synthesized. This indole derivative has a molecular architecture kindred to biologically relevant species and was obtained through five synthetic steps with an overall yield of 37% from the 2,2'-(phenylazanediyl)di(ethan-1-ol). The indole derivative was evaluated through docking assays using the main protease (SARS-CoV-2-M^pro) as a molecular target, which plays a key role in the replication process of this virus. Additionally, the indole derivative was evaluated as an inhibitor of the enzyme kallikrein 5 (KLK5), which is a serine protease that can be considered as an anticancer drug target.

Catalytic mechanism of the colistin resistance protein MCR-1

The mcr-1 gene encodes a membrane-bound Zn2+-metalloenzyme, MCR-1, which catalyses phosphoethanolamine transfer onto bacterial lipid A, making bacteria resistant to colistin, a last-resort antibiotic. Mechanistic understanding of this process remains incomplete. Here, we investigate possible catalytic pathways using DFT and ab initio calculations on cluster models and identify a complete two-step reaction mechanism. The first step, formation of a covalent phosphointermediate via transfer of phosphoethanolamine from a membrane phospholipid donor to the acceptor Thr285, is rate-limiting and proceeds with a single Zn2+ ion. The second step, transfer of the phosphoethanolamine group to lipid A, requires an additional Zn2+. The calculations suggest the involvement of the Zn2+ orbitals directly in the reaction is limited, with the second Zn2+ acting to bind incoming lipid A and direct phosphoethanolamine addition. The new level of mechanistic detail obtained here, which distinguishes these enzymes from other phosphotransferases, will aid in the development of inhibitors specific to MCR-1 and related bacterial phosphoethanolamine transferases.

Beyond generative models: superfast traversal, optimization, novelty, exploration and discovery (STONED) algorithm for molecules using SELFIES

Inverse design allows the generation of molecules with desirable physical quantities using property optimization. Deep generative models have recently been applied to tackle inverse design, as they possess the ability to optimize molecular properties directly through structure modification using gradients. While the ability to carry out direct property optimizations is promising, the use of generative deep learning models to solve practical problems requires large amounts of data and is very time-consuming. In this work, we propose STONED - a simple and efficient algorithm to perform interpolation and exploration in the chemical space, comparable to deep generative models. STONED bypasses the need for large amounts of data and training times by using string modifications in the SELFIES molecular representation. First, we achieve non-trivial performance on typical benchmarks for generative models without any training. Additionally, we demonstrate applications in high-throughput virtual screening for the design of drugs, photovoltaics, and the construction of chemical paths, allowing for both property and structure-based interpolation in the chemical space. Overall, we anticipate our results to be a stepping stone for developing more sophisticated inverse design models and benchmarking tools, ultimately helping generative models achieve wider adoption.

Theoretical and Mechanistic Validation of Global Kinetic Parameters of the Inactivation of GABA Aminotransferase by OV329 and CPP-115

((S)-3-Amino-(difluoromethylenyl)cyclopent-1-ene-1-carboxylic acid (OV329) is a recently discovered inactivator of γ-aminobutyric acid aminotransferase (GABA-AT), which has 10 times better inactivation efficiency than its predecessor, CPP-115, despite the only structural difference being an endocyclic double bond in OV329. Both compounds are mechanism-based enzyme inactivators (MBEIs), which inactivate GABA-AT by a similar mechanism. Here, a combination of a variety of computational chemistry tools and experimental methods, including quantum mechanical (QM) calculations, molecular dynamic simulations, progress curve analysis, and deuterium kinetic isotope effect (KIE) experiments, are utilized to comprehensively study the mechanism of inactivation of GABA-AT by CPP-115 and OV329 and account for their experimentally obtained global kinetic parameters k_inact and K_I. Our first key finding is that the rate-limiting step of the inactivation mechanism is the deprotonation step, and according to QM calculations and the KIE experiments, k_inact accurately represents the enhancement of the rate-limiting step for the given mechanism. Second, the present study shows that the widely used simple QM models do not accurately represent the geometric criteria that are present in the enzyme for the deprotonation step. In contrast, QM cluster models successfully represent both the ground state destabilization and the transition state stabilization, as revealed by natural bond orbital analysis. Furthermore, the globally derived K_I values for both of the inactivators represent the inhibitor constants for the initial binding complexes (K_d) and indicate the inactivator competition with the substrate according to progress curve analysis and the observed binding isotope effect. The configurational entropy loss accounts for the difference in K_I values between the inactivators. The approach we describe in this work can be employed to determine the validity of globally derived parameters in the process of MBEI optimization for given inactivation mechanisms.

MolFinder: an evolutionary algorithm for the global optimization of molecular properties and the extensive exploration of chemical space using SMILES

Here, we introduce a new molecule optimization method, MolFinder, based on an efficient global optimization algorithm, the conformational space annealing algorithm, and the SMILES representation. MolFinder finds diverse molecules with desired properties efficiently without any training and a large molecular database. Compared with recently proposed reinforcement-learning-based molecule optimization algorithms, MolFinder consistently outperforms in terms of both the optimization of a given target property and the generation of a set of diverse and novel molecules. The efficiency of MolFinder demonstrates that combinatorial optimization using the SMILES representation is a promising approach for molecule optimization, which has not been well investigated despite its simplicity. We believe that our results shed light on new possibilities for advances in molecule optimization methods.

Organic reactivity from mechanism to machine learning

As more data are introduced in the building of models of chemical reactivity, the mechanistic component can be reduced until 'big data' applications are reached. These methods no longer depend on underlying mechanistic hypotheses, potentially learning them implicitly through extensive data training. Reactivity models often focus on reaction barriers, but can also be trained to directly predict lab-relevant properties, such as yields or conditions. Calculations with a quantum-mechanical component are still preferred for quantitative predictions of reactivity. Although big data applications tend to be more qualitative, they have the advantage to be broadly applied to different kinds of reactions. There is a continuum of methods in between these extremes, such as methods that use quantum-derived data or descriptors in machine learning models. Here, we present an overview of the recent machine learning applications in the field of chemical reactivity from a mechanistic perspective. Starting with a summary of how reactivity questions are addressed by quantum-mechanical methods, we discuss methods that augment or replace quantum-based modelling with faster alternatives relying on machine learning.

Stochastic Modeling of Plant Virus Propagation with Biological Control

Plants are vital for man and many species. They are sources of food, medicine, fiber for clothes and materials for shelter. They are a fundamental part of a healthy environment. However, plants are subject to virus diseases. In plants most of the virus propagation is done by a vector. The traditional way of controlling the insects is to use insecticides that have a negative effect on the environment. A more environmentally friendly way to control the insects is to use predators that will prey on the vector, such as birds or bats. In this paper we modify a plant-virus propagation model with delays. The model is written using delay differential equations. However, it can also be expressed in terms of biochemical reactions, which is more realistic for small populations. Since there are always variations in the populations, errors in the measured values and uncertainties, we use two methods to introduce randomness: stochastic differential equations and the Gillespie algorithm. We present numerical simulations. The Gillespie method produces good results for plant-virus population models.

Synthesis, molecular docking, and in silico ADME/Tox profiling studies of new 1-aryl-5-(3-azidopropyl)indol-4-ones: Potential inhibitors of SARS CoV-2 main protease

The virus SARS CoV-2, which causes the respiratory infection COVID-19, continues its spread across the world and to date has caused more than a million deaths. Although COVID-19 vaccine development appears to be progressing rapidly, scientists continue the search for different therapeutic options to treat this new illness. In this work, we synthesized five new 1-aryl-5-(3-azidopropyl)indol-4-ones and showed them to be potential inhibitors of the SARS CoV-2 main protease (3CLpro). The compounds were obtained in good overall yields and molecular docking indicated favorable binding with 3CLpro. In silico ADME/Tox profile of the new compounds were calculated using the SwissADME and pkCSM-pharmacokinetics web tools, and indicated adequate values of absorption, distribution and excretion, features related to bioavailability. Moreover, low values of toxicity were indicated for these compounds. And drug-likeness levels of the compounds were also predicted according to the Lipinski and Veber rules.

myPresto/omegagene 2020: a molecular dynamics simulation engine for virtual-system coupled sampling

The molecular dynamics (MD) method is a promising approach for investigating the molecular mechanisms of microscopic phenomena. In particular, generalized ensemble MD methods can efficiently explore the conformational space with a rugged free-energy surface. However, the implementation and acquisition of technical knowledge for each generalized ensemble MD method are not straightforward for end-users. Here, we present a new version of the myPresto/omegagene software, which is an MD simulation engine tailored for a series of generalized ensemble methods, which are virtual-system coupled multicanonical MD (V-McMD), virtual-system coupled adaptive umbrella sampling (V-AUS), and virtual-system coupled canonical MD (VcMD). This program has been applied in several studies analyzing free-energy landscapes of a variety of molecular systems with all-atom simulations. The updated version provides new functionality for coarse-grained simulations powered by the hydrophobicity scale method. The software package includes a step-by-step tutorial document for enhanced conformational sampling of the poly-glutamine (poly-Q) oligomer expressed as a one-bead per residue model. The myPresto/omegagene software is freely available at the following URL: https://github.com/kotakasahara/omegagene under the Apache2 license.

Unraveling the SARS-CoV-2 Main Protease Mechanism Using Multiscale Methods

We present a detailed theoretical analysis of the reaction mechanism of proteolysis catalyzed by the main protease of SARS-CoV-2. Using multiscale simulation methods, we have characterized the interactions established by a peptidic substrate in the active site, and then we have explored the free energy landscape associated with the acylation and deacylation steps of the proteolysis reaction, characterizing the transition states of the process. Our mechanistic proposals can explain most of the experimental observations made on the highly similar ortholog protease of SARS-CoV. We point to some key interactions that may facilitate the acylation process and thus can be crucial in the design of more specific and efficient inhibitors of the main protease activity. In particular, from our results, the P1' residue can be a key factor to improve the thermodynamics and kinetics of the inhibition process.

Illuminating elite patches of chemical space

In the past few years, there has been considerable activity in both academic and industrial research to develop innovative machine learning approaches to locate novel, high-performing molecules in chemical space. Here we describe a new and fundamentally different type of approach that provides a holistic overview of how high-performing molecules are distributed throughout a search space. Based on an open-source, graph-based implementation [J. H. Jensen, Chem. Sci., 2019, 10, 3567-3572] of a traditional genetic algorithm for molecular optimisation, and influenced by state-of-the-art concepts from soft robot design [J. B. Mouret and J. Clune, Proceedings of the Artificial Life Conference, 2012, pp. 593-594], we provide an algorithm that (i) produces a large diversity of high-performing, yet qualitatively different molecules, (ii) illuminates the distribution of optimal solutions, and (iii) improves search efficiency compared to both machine learning and traditional genetic algorithm approaches.

Electronic structure benchmark calculations of CO2 fixing elementary chemical steps in RuBisCO using the projector-based embedding approach

Ribulose 1,5-bisphosphate carboxylase-oxygenase (RuBisCO) is the main enzyme involved in atmospheric carbon dioxide (CO₂ ) fixation in the biosphere. This enzyme catalyzes a set of five chemical steps that take place in the same active-site within magnesium (II) coordination sphere. Here, a set of electronic structure benchmark calculations have been carried out on a reaction path proposed by Gready et al. by means of the projector-based embedding approach. Activation and reaction energies for all main steps catalyzed by RuBisCO have been calculated at the MP2, SCS-MP2, CCSD, and CCSD(T)/aug-cc-pVDZ and cc-pVDZ levels of theory. The treatment of the magnesium cation with post-HF methods is explored to determine the nature of its involvement in the mechanism. With the high-level ab initio values as a reference, we tested the performance of a set of density functional theory (DFT) exchange-correlation (xc) functionals in reproducing the reaction energetics of RuBisCO carboxylase activity on a set of model fragments. Different DFT xc-functionals show large variation in activation and reaction energies. Activation and reaction energies computed at the B3LYP level are close to the reference SCS-MP2 results for carboxylation, hydration and protonation reactions. However, for the carbon-carbon bond dissociation reaction, B3LYP and other functionals give results that differ significantly from the ab initio reference values. The results show the applicability of the projector-based embedding approach to metalloenzymes. This technique removes the uncertainty associated with the selection of different DFT xc-functionals and so can overcome some of inherent limitations of DFT calculations, complementing, and potentially adding to modeling of enzyme reaction mechanisms with DFT methods.

Quantum-Chemistry Based Design of Halobenzene Derivatives With Augmented Affinities for the HIV-1 Viral G4/C16 Base-Pair

The HIV-1 integrase (IN) is a major target for the design of novel anti-HIV inhibitors. Among these, three inhibitors which embody a halobenzene ring derivative (HR) in their structures are presently used in clinics. High-resolution X-ray crystallography of the complexes of the IN-viral DNA transient complex bound to each of the three inhibitors showed in all cases the HR ring to interact within a confined zone of the viral DNA, limited to the highly conserved 5′CpA 3′/5′TpG 3′ step. The extension of its extracyclic CX bond is electron-depleted, owing to the existence of the “sigma-hole.” It interacts favorably with the electron-rich rings of base G₄. We have sought to increase the affinity of HR derivatives for the G₄/C₁₆ base pair. We thus designed thirteen novel derivatives and computed their Quantum Chemistry (QC) intermolecular interaction energies (ΔE) with this base-pair. Most compounds had ΔE values significantly more favorable than those of the HR of the most potent halobenzene drug presently used in clinics, Dolutegravir. This should enable the improvement in a modular piece-wise fashion, the affinities of halogenated inhibitors for viral DNA (vDNA). In view of large scale polarizable molecular dynamics simulations on the entirety of the IN-vDNA-inhibitor complexes, validations of the SIBFA polarizable method are also reported, in which the evolution of each ΔE(SIBFA) contribution is compared to its QC counterpart along this series of derivatives.

Metamorphic Proteins in Light of Anfinsen's Dogma

It is a common belief that metamorphic proteins challenge Anfinsen's thermodynamic hypothesis (or dogma). Here we argue against this view and aim to show that metamorphic proteins not only fulfill Anfinsen's dogma but also exhibit marginal stability comparable to that seen on biomolecules and macromolecular complexes. This work contributes to our general understanding of protein classification and may spur significant progress in our effort to analyze protein evolvability.

An Ab Initio Molecular Dynamics Study of the Hydrolysis Reaction of Sulfur Trioxide Catalyzed by a Formic Acid or Water Molecule

Ab initio molecular dynamics (AIMD) calculations have been performed to investigate the role of dynamical and steric effects in formic acid (FA) or H₂O-catalyzed gas phase hydrolysis of SO₃ to form sulfuric acid. This was done by colliding FA or H₂O with the SO₃-H₂O complex and the water dimer with the SO₃ molecule and analyzing the outcomes of 230 AIMD trajectories. Our calculations show that, within simulation times used, sulfuric acid is formed in 5% of FA collisions but is not produced when H₂O collides with the SO₃-H₂O complex or when the water dimer collides with the SO₃ molecule. We also find that FA collisions have about 2 times higher probability to form the prereactive complex than H₂O collisions. Moreover, our simulations show that the SO₃-H₂O-FA prereactive complex is more stable in time than the SO₃-H₂O-H₂O prereactive complex. These findings indicate that the FA-catalyzed mechanism is favored over the H₂O one when looking from the steric and dynamic effect point of view. Additionally, AIMD simulations starting from the optimized structure of the SO₃-H₂O-FA prereactive complex have been computed to qualitatively estimate the rate of the sulfuric acid formation. Collisional energy has been observed to promote sulfuric acid formation more effectively than thermal excitation.