In recent years, there has been progress in the development of machine learning and deep learning technologies in various fields, and a number of software packages have been released that can be implemented. Our research group has attempted to establish analysis methods using machine learning for various scientific data. In this paper, we will report on further developments such as prediction of lipophilicity of molecules, analysis of psalms data using natural language processing, and similarity calculation system of spectrum data.

The solvation free energies of cyclic dipeptides (CDPs) for a membrane were calculated to evaluate their membrane permeabilities. As a result, there were energy barriers between the membrane center and interface for each CDP. Furthermore, the profiles of solvation free energy depended on the amino acids that construct each CDP. These results indicate that each CDP changes its membrane permeability. Our results will be useful for designing cyclic peptides in the field of medium-molecular-weight drugs.

We implemented the picture change correction method for two-electron Coulomb interaction and density operator, which was based on the infinite-order two-component method with the local unitary transformation, to the GAMESS program. Numerical assessments for molecules containing heavy element confirmed the accuracies and efficiencies of the implementation. Furthermore, the comparison with several types of treatments indicated that whole picture change corrections of one- and two-electron operators and the density operator are necessary for accurate two-component density functional theory calculations.

This paper overviews the local structures and dynamics of imidazole (Im) in proton-conducting poly (vinylphosphonic acid)-Im (PVPA-xIm) composites investigated by density functional theory calculations and molecular dynamics simulations. These calculations propose the proton conduction mechanism of PVPA-xIm based on Grotthuss-type mechanism consisting the following processes: proton transfers from PVPA to Im and between Im, and the reorientation of Im in the hydrogen-bond Im cluster.

Motivated by a recent study on light-induced solar cell that Iridium-oxo complex over WO₃ electrode catalyze water oxidation reaction, we investigated the methane hydroxylation reaction by the same complex. Herein we focused on the closed-shell singlet state reaction due to the no-radical intermediate mechanism although this reaction is energetically unfavored reaction. Using intrinsic bond orbital analysis, we found that the C–H bond cleavage reaction proceed via the hydride transfer mechanism. From the view point of orbital interaction between σ orbital (C–H bond in methane) and π* orbital (Ir=O bond in Ir complex), we demonstrated that the reactivity of the methane hydroxylation reaction in the closed shell singlet state depends on the energy level of π* orbital.

In this study, we attempted to quantify the shape of various metastable structures of microphase separation of ABA triblock copolymers that can be observed in a computer simulation. For the quantification, we focused on the domains where the volume fraction of B (or A) is greater than a preset threshold, and defined a measure that indicates how the size of the largest connected component of such domains decreases as the threshold increases. The value of the measure was strongly correlated with the free energy of the metastable structures.

Silicon carbide (SiC) exhibits super low friction in water environments. It has been well noticed that this super low friction property is caused by the generated lubricating layers through friction-induced chemical reactions. However, the chemical reactions at the friction interface have not been well clarified because it is difficult to observe the tribological processes at the friction interface in experiments. In the present study, we simulated the friction processes of SiC/SiC in water environments using reactive molecular dynamics simulations. It is interesting to observe that the hydrolysis reactions of Si-C and Si-Si bonds occurred, and however, the chemical reactions of C-C bonds hardly occur, indicating that Si atoms are preferentially oxidized and C atoms are not. The oxidation of Si atoms in SiC generates colloidal silica lubricating layers at the friction interface, being responsible for the super low friction of SiC.

We have been developing the ABINIT-MP program for fragment molecular orbital (FMO) calculations. In this report, improvements in the latest release Ver. 1 Rev. 22 and demonstrative applications on the Fugaku and ITO supercomputers are summarized.

Anisotropic low-thermal expansion of β-LiAlSiO₄ crystal was reproduced by MD calculation applying a potential model of a spherically symmetric two-body interatomic interaction. From the correlation between the Si-O-Al angle and the Si-Al distance in the c-axis direction, the expansion of a-axis and the contract of c-axis brings the split of Si-O-Al angle distribution. At the same time, the stable position of Li⁺ ion sites also changed. This suggests the cation position also important in the mechanism of low-thermal expansion observed in silicate materials.

In order to accelerate the conformational search of peptides, a coarse-grained model in which amino acid residues were treated with an amino acid pair interaction potential was investigated. To treat both natural and artificial types of amino acids and staples, new interaction potentials were developed by umbrella sampling. For the new potential, the effect of optical isomer was investigated with the threonine potential. As a result, the potential of allo-threonine was different from that of L-threonine.

Ni-ZSM-5 is a promising catalyst for the conversion of methane to methanol. In many methane oxidation catalysts, the problem is that the oxidation reaction continues after methane oxidation (over-oxidation). The ease of over-oxidation is an important factor in evaluating the performance of methane oxidation catalysts. In the present study, we performed density functional theory (DFT) calculation to evaluate the two reaction pathways for the oxidation of methanol to formaldehyde by Ni-ZSM-5. DFT calculations indicated that the activation energies for the C-H bond cleavage are more than 40 kcal/mol in the reaction pathways. After methane oxidation the second step reaction corresponding to the methanol oxidation by [Ni-O-Ni]²⁺ is unlikely to occur.

The time-series clustering method is one of unsupervised machine learning techniques that classify time-series data. In this article, we applied three methods to the clustering analysis for 200 molecular dynamics (MD) trajectories of human adult hemoglobin (HbA), and have reported their clustering performances for detecting the T-R state transition trajectories (Traj_T-R). By compared with their silhouette indices, we have discussed the proper clustering conditions.

Single-molecule junctions have unique electron transport properties that are not found in solids or polymers. Graph theoretical approaches have been used to predict whether conductivity in single-molecule junctions. In this paper, the change in single-molecule conductivity when carbon atoms in alternant hydrocarbons are replaced with heteroatoms is investigated by using the graph-theoretic path-counting method. The results predicted by graph theory are consistent with those obtained by using nonequilibrium Green's function combined with the Hückel method.

Molecular dynamics (MD) simulations were used to analyze the effect of the odorant molecule eugenol on the olfactory receptor protein mOR-EG and the dynamic correlation between amino acid residues around the binding site due to ligand binding. When eugenol hydrogen-bonded with Ser 113 of mOR-EG, the dynamic correlation with amino acid residues 213-218 on the adjacent α-helix increased within 10 ps. In particular, the correlation between Ser 113 and Leu 217 showed characteristic oscillations between 10 ps and 30 ps. The correlation with Leu 217 shifted toward residues closer to the G protein over time. Correlation with amino acids on the other side of the membrane was also observed, suggesting the contribution of collective protein motion

Tapping test is an inspection method that determines the presence or absence of abnormality in a specimen based on the difference in the sound when tapping a material. This method is used to inspect buildings and railway vehicles. It is considered that this method can be used for quality evaluation of cast iron. However, although the tapping test has the advantage of being able to be performed non-destructively and simply, it also has the disadvantage of requiring a worker who can distinguish sounds. In order to solve this problem, we introduced a neural network and studied whether it is possible to judge the quality of cast iron by learning the tapping sound of cast iron.

The graphite structure of flake graphite cast iron is classified into 5 types of A,B,C,D,E or 6 types including chill. In usually, type A is good quality and type D is poor quality. As the proportion of D-type graphite increases, the tensile strength decreases.

Inspection of graphite structure is one of the quality evaluations. However, there is no method for numerically determining the graphite structure of flake graphite cast iron. In this study, we created an AI that determines 6 types of graphite structure from A to chill using a Convolutional Neural Network (CNN).The created AI judged the graphite structure photographs that not used for learning with an average correct answer rate of 97.5%. In addition, when the graphite structure of one sample was analyzed by AI, it was confirmed numerically that the graphite structure deteriorated.

Block copolymers with repulsive sub-chains form a variety of domain structures called microphase-separated structures. Most microphase-separated structures have a periodic patterned structure while in a stable state. On the other hand, metastable states with an acyclic structure and sufficiently long lifetimes are often observed around the configuration space of the periodic structures. In this study, we focused on the double gyroid structure—a type of microphase-separated structure—and built a model to predict free energy using topological data analysis (TDA), which has recently proven useful for analyzing acyclic compound structures such as the amorphous structure of glass. The model obtained can predict the free energy of the metastable state of the double-gyroid structure using persistent homology, which is one of the features of TDA.

Organic solar cells are flexible elements that offer benefits such as low cost, light weight, and applicability. For this study, quantum chemistry calculations of cells using conductive polymers were performed. Poly (3,4-ethylenedioxythiophene)-poly (styrenesulfonate)(PEDOT:PSS), phthalocyanine and fullerene C₆₀ were used as semiconductors. Generation of carriers (conduction electrons and holes) was expected in both the ground state and the excited state. When MgAg and Al were selected for the electrodes, the energy diagram was found to have an ideal step structure. The organic solar cell was designed from the energy diagram of the whole layer that constituted the solar cell.

The reaction process and steric situations of novel basic and chiral catalyst Diels-Alder reactions by Kagan et al. were speculated by IRC of PM7 simulation for the three molecules reactions clearly. The addition reactions of enolic dienes (1) with dienophiles (2) by amines (3) such as (S)-(+)-prolinol / (R)-(−)-prolinol proceeded via lower energy reaction complexes (RC) and transition states (TS) of two steps. The steric shapes by IRC (Figure 2 ∼ 6) showed the clear interactions between the reaction points, and of OH with amine moieties in the 1·3, 1·3·2 and TS complexes, to give high stereoselective adducts. IRC of some reactions also guesses right the Michael reaction selectivity. The handy PM7 simulation is recommended for usual chemical growth.