蛋白质-药物分子体系第一性原理光谱结构示范性数据集

我们选取大型药物分子数据库Chembl中获得FDA批准的活性小分子数据、和蛋白结构数据库RCSB选取新冠病毒的结构蛋白，主蛋白酶等一系列具有标志性意义的蛋白结构，同时选取这些蛋白的抗体蛋白结构以及和新冠类似的流感蛋白结构，和动力学构像结构等数据，利用上述药物小分子结构和标志性蛋白结构，并基于经验自由能力场和快速遗传算法的搜索蛋白质-配体小分子最稳定的结合构象。基于上述结构数据，我们结合坐标信息和电子密度信息的kmeans智能分块算法，将大分子分成小块结构，基于第一性原理高通量计算光谱数据并构建数据集。

We collected FDA-approved active small-molecule data from the large-scale drug molecular database ChEMBL, as well as a series of landmark protein structures including SARS-CoV-2 structural proteins and main proteases retrieved from the RCSB Protein Data Bank. Furthermore, we gathered data such as antibody protein structures of these proteins, influenza protein structures homologous to SARS-CoV-2, and dynamic conformational structures. Leveraging the aforementioned small-molecule drug and landmark protein structures, we employed the empirical free energy force field and fast genetic algorithm to identify the most stable binding conformations between proteins and their small-molecule ligands. Based on these structural data, we utilized a K-means intelligent partitioning algorithm that integrates coordinate and electron density information to divide large macromolecules into smaller structural blocks. Subsequently, we calculated spectroscopic data via high-throughput first-principles calculations and constructed the final dataset.

搜集汇总

背景与挑战

背景概述

该数据集是一个基于第一性原理计算的蛋白质-药物分子体系光谱结构数据集，包含FDA批准的活性小分子和新冠病毒相关蛋白结构，通过智能分块算法和高通量计算构建，适用于量子化学和分子生物学研究。