DFRscore: Deep Learning-Based Scoring of Synthetic Complexity with Drug-Focused Retrosynthetic Analysis for High-Throughput Virtual Screening.

Recently emerging generative AI models enable us to produce a vast number of compounds for potential applications. While they can provide novel molecular structures, the synthetic feasibility of the generated molecules is often questioned. To address this issue, a few recent studies have attempted to use deep learning models to estimate the synthetic accessibility of many molecules rapidly. However, retrosynthetic analysis tools used to train the models rely on reaction templates automatically extracted from a large reaction database that are not domain-specific and may exhibit low chemical correctness. To overcome this limitation, we introduce DFRscore (Drug-Focused Retrosynthetic score), a deep learning-based approach for a more practical assessment of synthetic accessibility in drug discovery. The DFRscore model is trained exclusively on drug-focused reactions, providing a predicted number of minimally required synthetic steps for each compound. This approach enables practitioners to filter out compounds that do not meet their desired level of synthetic accessibility at an early stage of high-throughput virtual screening for accelerated drug discovery. The proposed strategy can be easily adapted to other domains by adjusting the synthesis planning setup of the reaction templates and starting materials.

Deep Learning-Based Chemical Similarity for Accelerated Organic Light-Emitting Diode Materials Discovery.

Thermally activated delayed fluorescence (TADF) material has attracted great attention as a promising metal-free organic light-emitting diode material with a high theoretical efficiency. To accelerate the discovery of novel TADF materials, computer-aided material design strategies have been developed. However, they have clear limitations due to the accessibility of only a few computationally tractable properties. Here, we propose TADF-likeness, a quantitative score to evaluate the TADF potential of molecules based on a data-driven concept of chemical similarity to existing TADF molecules. We used a deep autoencoder to characterize the common features of existing TADF molecules with common chemical descriptors. The score was highly correlated with the four essential electronic properties of TADF molecules and had a high success rate in large-scale virtual screening of millions of molecules to identify promising candidates at almost no cost, validating its feasibility for accelerating TADF discovery. The concept of TADF-likeness can be extended to other fields of materials discovery.

Gaussian-Approximated Poisson Preconditioner for Iterative Diagonalization in Real-Space Density Functional Theory.

Various real-space methods optimized on massive parallel computers have been developed for efficient large-scale density functional theory (DFT) calculations of materials and biomolecules. The iterative diagonalization of the Hamiltonian matrix is a computational bottleneck in real-space DFT calculations. Despite the development of various iterative eigensolvers, the absence of efficient real-space preconditioners has hindered their overall efficiency. An efficient preconditioner must satisfy two conditions: appropriate acceleration of the convergence of the iterative process and inexpensive computation. This study proposed a Gaussian-approximated Poisson preconditioner (GAPP) that satisfied both conditions and was suitable for real-space methods. A low computational cost was realized through the Gaussian approximation of a Poisson Green's function. Fast convergence was achieved through the proper determination of Gaussian coefficients to fit the Coulomb energies. The performance of GAPP was evaluated for several molecular and extended systems, and it showed the highest efficiency among the existing preconditioners adopted in real-space codes.

Circularly Polarized Light Can Override and Amplify Asymmetry in Supramolecular Helices.

Circularly polarized light (CPL) is an inherently chiral entity and is considered one of the possible deterministic signals that led to the evolution of homochirality. While accumulating examples indicate that chirality beyond the molecular level can be induced by CPL, not much is yet known about circumstances where the spin angular momentum of light competes with existing molecular chiral information during the chirality induction and amplification processes. Here we present a light-triggered supramolecular polymerization system where chiral information can both be transmitted and nonlinearly amplified in a "sergeants-and-soldiers" manner. While matching handedness with CPL resulted in further amplification, we determined that opposite handedness could override molecular information at the supramolecular level when the enantiomeric excess was low. The presence of a critical chiral bias suggests a bifurcation point in the homochirality evolution under random external chiral perturbation. Our results also highlight opportunities for the orthogonal control of supramolecular chirality decoupled from molecular chirality preexisting in the system.

Neural network-based pseudopotential: development of a transferable local pseudopotential.

Transferable local pseudopotentials (LPPs) are essential for fast quantum simulations of materials. However, various types of LPPs suffer from low transferability, especially since they do not consider the norm-conserving condition. Here we propose a novel approach based on a deep neural network to produce transferable LPPs. We introduced a generalized Kerker method expressed with the deep neural network to represent the norm-conserving pseudo-wavefunctions. Its unique feature is that all necessary conditions of pseudopotentials can be explicitly considered in terms of a loss function. Then, it can be minimized using the back-propagation technique just with single point all-electron atom data. To assess the transferability and accuracy of the neural network-based LPPs (NNLPs), we carried out density functional theory calculations for the s- and p-block elements of the second to the fourth periods. The NNLPs outperformed other types of LPPs in both atomic and bulk calculations for most elements. In particular, they showed good transferability by predicting various properties of bulk systems including binary alloys with higher accuracy than LPPs tailored to bulk data.

Morphology Transformation of Foldamer Assemblies Triggered by Single Oxygen Atom on Critical Residue Switch.

The synthesis of morphologically well-defined peptidic materials via self-assembly is challenging but demanding for biocompatible functional materials. Moreover, switching morphology from a given shape to other predictable forms by molecular modification of the identical building block is an even more complicated subject because the self-assembly of flexible peptides is prone to diverge upon subtle structural change. To accomplish controllable morphology transformation, systematic self-assembly studies are performed using congener short ß-peptide foldamers to find a minimal structural change that alters the self-assembled morphology. Introduction of oxygen-containing ß-amino acid (ATFC) for subtle electronic perturbation on hydrophobic foldamer induces a previously inaccessible solid-state conformational split to generate the most susceptible modification site for morphology transformation of the foldamer assemblies. The site-dependent morphological switching power of ATFC is further demonstrated by dual substitution experiments and proven by crystallographic analyses. Stepwise morphology transformation is shown by modifying an identical foldamer scaffold. This study will guide in designing peptidic molecules from scratch to create complex and biofunctional assemblies with nonspherical shapes.

Molecular Generative Model Based on an Adversarially Regularized Autoencoder.

Deep generative models are attracting great attention as a new promising approach for molecular design. A variety of models reported so far are based on either a variational autoencoder (VAE) or a generative adversarial network (GAN), but they have limitations such as low validity and uniqueness. Here, we propose a new type of model based on an adversarially regularized autoencoder (ARAE). It basically uses latent variables like VAE, but the distribution of the latent variables is estimated by adversarial training like in GAN. The latter is intended to avoid both the insufficiently flexible approximation of posterior distribution in VAE and the difficulty in handling discrete variables in GAN. Our benchmark study showed that ARAE indeed outperformed conventional models in terms of validity, uniqueness, and novelty per generated molecule. We also demonstrated a successful conditional generation of drug-like molecules with ARAE for the control of both cases of single and multiple properties. As a potential real-world application, we could generate epidermal growth factor receptor inhibitors sharing the scaffolds of known active molecules while satisfying drug-like conditions simultaneously.

Graph theory-based reaction pathway searches and DFT calculations for the mechanism studies of free radical-initiated peptide sequencing mass spectrometry (FRIPS MS): a model gas-phase reaction of GGR tri-peptide.

Graph theory-based reaction pathway searches (ACE-Reaction program) and density functional theory calculations were performed to shed light on the mechanisms for the production of [an + H]+, xn+, yn+, zn+, and [yn + 2H]+ fragments formed in free radical-initiated peptide sequencing (FRIPS) mass spectrometry measurements of a small model system of glycine-glycine-arginine (GGR). In particular, the graph theory-based searches, which are rarely applied to gas-phase reaction studies, allowed us to investigate reaction mechanisms in an exhaustive manner without resorting to chemical intuition. As expected, radical-driven reaction pathways were favorable over charge-driven reaction pathways in terms of kinetics and thermodynamics. Charge- and radical-driven pathways for the formation of [yn + 2H]+ fragments were carefully compared, and it was revealed that the [yn + 2H]+ fragments observed in our FRIPS MS spectra originated from the radical-driven pathway, which is in contrast to the general expectation. The acquired understanding of the FRIPS fragmentation mechanism is expected to aid in the interpretation of FRIPS MS spectra. It should be emphasized that graph theory-based searches are powerful and effective methods for studying reaction mechanisms, including gas-phase reactions in mass spectrometry.

Holistic Approach to the Mechanism Study of Thermal Degradation of Organic Light-Emitting Diode Materials.

The design of stable organic light-emitting diode materials is the key to long lifetime displays under various stressful conditions. Elucidating the degradation mechanism of the materials at the molecular level provides useful information for securing high stability. Previous works based on experiments or computations disclosed only a part of the whole degradation process. Here, we propose a holistic approach to the systematic analysis of the degradation mechanism by combining experimental mass analysis and computation in a semi-automated fashion. The mass analysis identifies molecular weights of feasible products from degradation reactions. Then, the computational analysis goes through initiation, propagation, and termination phases. The initiation phase determines radical fragments and reactive sites, triggering the propagation process. In the propagation phase, we subsequently perform intermediate sampling, reaction network construction, and kinetic analysis. As a proof of concept, this approach was applied to the thermal degradation problem during the sublimation purification process. Two major pathways were successfully elucidated with full atomistic details.

ACE-Molecule: An open-source real-space quantum chemistry package.

ACE-Molecule (advanced computational engine for molecules) is a real-space quantum chemistry package for both periodic and non-periodic systems. ACE-Molecule adopts a uniform real-space numerical grid supported by the Lagrange-sinc functions. ACE-Molecule provides density functional theory (DFT) as a basic feature. ACE-Molecule is specialized in efficient hybrid DFT and wave-function theory calculations based on Kohn-Sham orbitals obtained from a strictly localized exact exchange potential. It is open-source oriented calculations with a flexible and convenient development interface. Thus, ACE-Molecule can be improved by actively adopting new features from other open-source projects and offers a useful platform for potential developers and users. In this work, we introduce overall features, including theoretical backgrounds and numerical examples implemented in ACE-Molecule.

Homochiral Supramolecular Thin Film from Self-Assembly of Achiral Triarylamine Molecules by Circularly Polarized Light.

Here, we report the formation of homochiral supramolecular thin film from achiral molecules, by using circularly polarized light (CPL) only as a chiral source, on the condition that irradiation of CPL does not induce a photochemical change of the achiral molecules. Thin films of self-assembled structures consisting of chiral supramolecular fibrils was obtained from the triarylamine derivatives through evaporation of the self-assembled triarylamine solution. The homochiral supramolecular helices with the desired handedness was achieved by irradiation of circularly polarized visible light during the self-assembly process, and the chiral stability of supramolecular self-assembled product was achieved by photopolymerization of the diacetylene moieties at side chains of the building blocks, with irradiation of circularly polarized ultraviolet light. This work provides a novel methodology for the generation of homochiral supramolecular thin film from the corresponding achiral molecules.

Predicting Drug-Target Interaction Using a Novel Graph Neural Network with 3D Structure-Embedded Graph Representation.

We propose a novel deep learning approach for predicting drug-target interaction using a graph neural network. We introduce a distance-aware graph attention algorithm to differentiate various types of intermolecular interactions. Furthermore, we extract the graph feature of intermolecular interactions directly from the 3D structural information on the protein-ligand binding pose. Thus, the model can learn key features for accurate predictions of drug-target interaction rather than just memorize certain patterns of ligand molecules. As a result, our model shows better performance than docking and other deep learning methods for both virtual screening (AUROC of 0.968 for the DUD-E test set) and pose prediction (AUROC of 0.935 for the PDBbind test set). In addition, it can reproduce the natural population distribution of active molecules and inactive molecules.

Performance of ACE-Reaction on 26 Organic Reactions for Fully Automated Reaction Network Construction and Microkinetic Analysis.

Accurate analysis of complex chemical reaction networks is necessary for reliable prediction of reaction mechanism. Though quantum chemical methods provide a desirable accuracy, large computational costs are unavoidable as considering numerous reaction pathways on the networks. We proposed a graph-theoretic approach combined with chemical heuristics (named ACE-Reaction) in previous work [ Chem. Sci. 2018 , 9 , 825 ], which automatically and rapidly finds out the most essential part of reaction networks just from reactants and products, and here we extended it by incorporating a stochastic approach for microkinetic modeling. To show its performance and broad applicability, we applied it to 26 organic reactions, which include 16 common functional groups. As a result, we could demonstrate that ACE-Reaction successfully found the accepted mechanism of all reactions, most within a few hours on a single workstation, and additional microkinetic modeling automatically discovered new competitive paths as well as a major path.

Feasibility of Activation Energy Prediction of Gas-Phase Reactions by Machine Learning.

Machine learning based on big data has emerged as a powerful solution in various chemical problems. We investigated the feasibility of machine learning models for the prediction of activation energies of gas-phase reactions. Six different models with three different types, including the artificial neural network, the support vector regression, and the tree boosting methods, were tested. We used the structural and thermodynamic properties of molecules and their differences as input features without resorting to specific reaction types so as to maintain the most general input form for broad applicability. The tree boosting method showed the best performance among others in terms of the coefficient of determination, mean absolute error, and root mean square error, the values of which were 0.89, 1.95, and 4.49 kcal mol-1 , respectively. Computation time for the prediction of activation energies for 2541 test reactions was about one second on a single computing node without using accelerators.

Efficient structural elucidation of microhydrated biomolecules through the interrogation of hydrogen bond networks.

Microhydration of biomolecules is an important structural factor associated with their biological processes. However, there is no general way to elucidate stable hydrated structures even for simple amino acids because of the high complexity of chemical space increasing rapidly with the number of water molecules. Here, we propose a very efficient computational method to selectively sample the most stable structures of microhydrated molecules. The key idea is to utilize the unique structural patterns of H-bond networks obtained from their energetic features, i.e. their tendency to form more H-bonds. As a proof of concept, we could identify the new global minima of glycine·10(H2O) and for the first time, we found the minimum number of water molecules required to stabilize the zwitterionic form of tyrosine. Furthermore, the most stable structures of hydrated glycine and tyrosine indeed had common features, which were consistent with the X-ray data of proteins in water.

Fragment-orbital tunneling currents and electronic couplings for analysis of molecular charge-transfer systems.

In theoretical charge-transfer research, calculation of the electronic coupling element is crucial for examining the degree of the electronic donor-acceptor interaction. The tunneling current (TC), representing the magnitudes and directions of electron flow, provides a way of evaluating electronic couplings, along with the ability of visualizing how electrons flow in systems. Here, we applied the TC theory to π-conjugated organic dimer systems, in the form of our fragment-orbital tunneling current (FOTC) method, which uses the frontier molecular-orbitals of system fragments as diabatic states. For a comprehensive test of FOTC, we assessed how reasonable the computed electronic couplings and the corresponding TC densities are for the hole- and electron-transfer databases HAB11 and HAB7. FOTC gave 12.5% mean relative unsigned error with regard to the high-level ab initio reference. The shown performance is comparable with that of fragment-orbital density functional theory, which gave the same error by 20.6% or 13.9% depending on the formulation. In the test of a set of nucleobase π stacks, we showed that the original TC expression is also applicable to nondegenerate cases under the condition that the overlap between the charge distributions of diabatic states is small enough to offset the energy difference. Lastly, we carried out visual analysis on the FOTC densities of thiophene dimers with different intermolecular alignments. The result depicts an intimate topological connection between the system geometry and electron flow. Our work provides quantitative and qualitative grounds for FOTC, showing it to be a versatile tool in characterization of molecular charge-transfer systems.

Effects of the locality of a potential derived from hybrid density functionals on Kohn-Sham orbitals and excited states.

Density functional theory (DFT) has been an essential tool for electronic structure calculations in various fields. In particular, its hybrid method including the Hartree-Fock (HF) exchange term remarkably improves the reliability of DFT for chemical applications and computational material design. There are two different types of exchange-correlation potential that can be derived from hybrid functionals. The conventional approach adopts a non-multiplicative potential including the non-local HF exchange operator. Herein, we propose to use a local multiplicative potential as an alternative for accurate excited state calculations. We show that such a local potential can be derived from existing global hybrid functionals using the optimized effective potential method. As a proof-of-concept, we chose PBE0 and investigated its performance for the Caricato benchmark set. Unlike the conventional one, the local potential produced orbital energy gaps with no strong dependence on the mixing ratio as a good approximation for optical excitations. Furthermore, its time-dependent DFT resulted in a surprisingly small mean absolute error even with a local density approximation kernel, surpassing all reported values with various popular functionals. In particular, most excitations were dictated by single orbital transitions due to physically meaningful virtual orbitals, which is beneficial to clear interpretations in the molecular orbital picture.

Performance of heterogeneous computing with graphics processing unit and many integrated core for hartree potential calculations on a numerical grid.

We investigated the performance of heterogeneous computing with graphics processing units (GPUs) and many integrated core (MIC) with 20 CPU cores (20×CPU). As a practical example toward large scale electronic structure calculations using grid-based methods, we evaluated the Hartree potentials of silver nanoparticles with various sizes (3.1, 3.7, 4.9, 6.1, and 6.9 nm) via a direct integral method supported by the sinc basis set. The so-called work stealing scheduler was used for efficient heterogeneous computing via the balanced dynamic distribution of workloads between all processors on a given architecture without any prior information on their individual performances. 20×CPU + 1GPU was up to ∼1.5 and ∼3.1 times faster than 1GPU and 20×CPU, respectively. 20×CPU + 2GPU was ∼4.3 times faster than 20×CPU. The performance enhancement by CPU + MIC was considerably lower than expected because of the large initialization overhead of MIC, although its theoretical performance is similar with that of CPU + GPU. © 2016 Wiley Periodicals, Inc.

Outstanding performance of configuration interaction singles and doubles using exact exchange Kohn-Sham orbitals in real-space numerical grid method.

To assess the performance of multi-configuration methods using exact exchange Kohn-Sham (KS) orbitals, we implemented configuration interaction singles and doubles (CISD) in a real-space numerical grid code. We obtained KS orbitals with the exchange-only optimized effective potential under the Krieger-Li-Iafrate (KLI) approximation. Thanks to the distinctive features of KLI orbitals against Hartree-Fock (HF), such as bound virtual orbitals with compact shapes and orbital energy gaps similar to excitation energies; KLI-CISD for small molecules shows much faster convergence as a function of simulation box size and active space (i.e., the number of virtual orbitals) than HF-CISD. The former also gives more accurate excitation energies with a few dominant configurations than the latter, even with many more configurations. The systematic control of basis set errors is straightforward in grid bases. Therefore, grid-based multi-configuration methods using exact exchange KS orbitals provide a promising new way to make accurate electronic structure calculations.

Supersampling method for efficient grid-based electronic structure calculations.

The egg-box effect, the spurious variation of energy and force due to the discretization of continuous space, is an inherent vexing problem in grid-based electronic structure calculations. Its effective suppression allowing for large grid spacing is thus crucial for accurate and efficient computations. We here report that the supersampling method drastically alleviates it by eliminating the rapidly varying part of a target function along both radial and angular directions. In particular, the use of the sinc filtering function performs best because as an ideal low pass filter it clearly cuts out the high frequency region beyond allowed by a given grid spacing.