Dataset associated with the manuscript titled "Capturing non-local through-bond effects when fragmenting molecules for quantum chemical torsion scans"

The dataset is a part of the supporting information of the fragmenter manuscript, which also describes how this dataset was generated in more detail.

This dataset includes all fragments generated from 730 FDA approved molecule without fragmenting rings.

Scripts used to generate this set live here.

1. Filter DrugBank with the following criteria:
    1. FDA approved
    2. Largest ring size has less than 14 heavy atoms
    3. Smallest ring size has at least 3 heavy atoms
    4. Molecule has less than 10 rotatable bonds
    5. Molecule must have at least one aromatic ring
    6. Molecule has only one connected component
2. Enumerate tautomers at physiological pH (this step generated 1243 molecules)
3. Exhaustive fragmentation without fragmenting rings (~300,000 fragments)
4. Generate conformations for each fragment using Omega
5. Calculate AM1 WBO for each conformation

本数据集属于该片段化研究论文补充材料的一部分，其中亦详细阐述了本数据集的生成流程。

本数据集涵盖从730种经美国食品药品监督管理局（FDA）批准的分子中生成的全部未断裂环结构的片段。

用于生成该数据集的脚本可在此处获取。

1. 按照以下标准对DrugBank数据库进行筛选：
1. 分子经FDA批准
2. 最大环的重原子数少于14个
3. 最小环的重原子数不少于3个
4. 分子的可旋转键数量少于10个
5. 分子至少含有一个芳香环
6. 分子仅包含一个连通组分
2. 在生理pH条件下枚举互变异构体（此步骤共生成1243个分子）
3. 进行穷举式片段化且不破坏环结构，最终得到约300,000个片段
4. 使用Omega工具为每个片段生成构象
5. 为每个构象计算AM1 WBO