A novel periodic table was developed. In this table, atomic orbital(s) characterizing each element is assigned and displayed either in the form of electron cloud-like representation or isosurface representation. From this table we can recognize the periodicity from the pictures of these orbitals having wave character, as if we could hear the sound of periodicity.
The molecular dynamics program: LAMMPS was implemented with highly efficient parallelization, and even super computer K computer demonstrated its high performance. In order to further speed up the molecular dynamics program LAMMPS, we tried to do it by converting the random number generation routine into Single Instruction Multi Data (SIMD / vector). We newly implemented speeding up of the difficult part of parallelization by SIMD, the part where large-scale parallel processing such as random number generation is usually not possible. In particular, a performance improvement of about 46% was observed overall due to the improvement of the implementation of the random number generation. It looks like there must still be some potential speedups.
We theoretically study the electronic structure of group-IV nanosheets, (graphene, silicene and germanene) having several kinds of imperfections such as atomic vacancies and heteroatoms replacement. We determine the atomistic geometries including above imperfections by employing the first-principles cluster calculation, and then explore the band structures by the supercell method. A single- and double-atom vacancies from graphene does not break the molecular planarity. Similarly, atomic vacancies in silicene and germanene hardly changes the non-planar corrugation found in their original molecular planes. Although the pair replacement by boron and nitrogen compensates the number of (pseudo-)π electrons, the difference in their on-site energies opens a small band gap at the “Dirac” point.
In quantitative structure-activity relationship and quantitative structure-physical relationship quantitatively, regression models are constructed activities and properties y, and molecular descriptors x for compounds. To improve predictive performance of models, multiple sub-models are constructed and a final y-value is predicted by integrating y-values predicted with sub-models in ensemble learning. Although it was confirmed that predictive performance improved by considering the applicability domain (AD) of each sub-model and by using only the sub-models inside AD, ADs cannot be compared between sub-datasets with different x. It was impossible to predict a y-value by selecting and weighting sub-models for a new sample. In this study, we focused on the similarity-weighted root-mean-square distance (wRMSD), which is an index of AD, and developed wRMSD-based AD considering ensemble learning (WEL), an ensemble learning method based on wRMSD. Since wRMSD is represented as the scale of y, AD can be compared between sub-models with different x, and thus, it is possible to predict a y-value, weighting sub-models having low wRMSD-values, which means high reliability of prediction, for a new sample. It was confirmed that AD was enlarged and predictive performance improved by using WEL compared to the conventional ensemble learning method through data analysis using three datasets of compounds for which water solubility, toxicity and pharmacological activity were measured. Python code for WEL is available at https://github.com/hkaneko1985/wel.
Gaussian and GAMESS, which are the codes of the ab initio molecular orbital method, can be used simply by specifying the name of the basis set such as STO-3G. These make a difference in the result in spite of using the same name. It is serious because the results have to be the same using the same input. In this paper, we call attention to the use of both programs with STO-3G for showing hydrides of the third period elements (NaH, MgH2, AlH3, SiH4, PH3, H2S, HCl) to alert users who are conducting research using two programs and the calculated results were comparable. As a result, except for HCl, the total energy and the energy of HOMO were lower in GAMESS than in Gaussian. This is due to the difference in the spatial spread of the 3s orbital and the 3p orbital of the third period element. The Mulliken charge is smaller in GAMESS than in Gaussian. When using some other programs, it is important to examine whether the input data can describe the electronic state intended by the user.