PubChemQC PM6
收藏资源简介:
PubChemQC PM6数据集是由理化学研究所的研究人员创建,包含了2.21亿个分子的优化几何结构和电子性质。该数据集基于PM6方法计算,覆盖了PubChem化合物数据库中92.9%的分子。数据集内容丰富,包括中性、阳离子、阴离子和自旋翻转电子态的分子。创建过程中,研究人员使用了SMILES和InChI编码来处理分子数据。该数据集广泛应用于有机薄膜太阳能电池、电致发光材料、有机非线性光学材料、分子传感器和新药设计等领域,旨在通过量子化学计算提供高质量的训练数据,以促进机器学习在化学领域的应用。
The PubChemQC PM6 Dataset was created by researchers at RIKEN. It contains optimized geometric structures and electronic properties of 221 million molecules. Computed using the PM6 semiempirical quantum chemistry method, the dataset covers 92.9% of the molecules in the PubChem Compound Database. The dataset encompasses diverse molecular types, including neutral, cationic, anionic, and spin-flip electronic state molecules. During its construction, researchers utilized SMILES and InChI encodings to process molecular data. This dataset is widely applied in research fields such as organic thin-film solar cells, electroluminescent materials, organic nonlinear optical materials, molecular sensors, and novel drug design. It aims to provide high-quality training data via quantum chemical calculations, so as to facilitate the application of machine learning in the field of chemistry.



