MBC and ECBL Libraries: outstanding tools for drug discovery

<strong>UPDATE</strong>. New in this revision: python scripts to process DBs and calculate the percentage of molecules which pass the Veber and Ghose filters. - <em>Veber_filter.py</em> ad <em>Ghose filter.py </em>(data presented in <strong>Table 1 </strong>of the MS). Data and scripts to reproduce all the graphics reported in the Manuscript entitled: "MBC and ECBL Libraries: outstanding tools for drug discovery". <strong>List of analyzed DBs:</strong> MBC2016 (Total entries: 1,096 cmpds; 7.39% excluded from properties analysis - QikProp failure). MBC2022 (Total entries: 2,577 cmpds; 3.14% excluded from properties analysis - QikProp failure). ECBL (Total entries: 101,021 cmpds; 0.20% excluded from properties analysis - QikProp failure). ChEMBL v.31 (Total entries 1,908,325 cmpds; 2.97% excluded from properties analysis - QikProp failure). DrugBank v.5.0 (Total entries 10,981 cmpds; 4.13% excluded from properties analysis - QikProp failure). ZINC20 (Total entries 10,723,360 cmpds; 0.61% excluded from properties analysis - QikProp failure). <strong>Files:</strong> <em>QikProp_properties.docx</em>: doc file containing the full list of QikProp properties calculated for each analyzed DB. <em>DATA_comparison.xlsx</em>: excel file containing data used to reproduce plots in <strong>Figure 4</strong> of the MS. <em>Murcko_scaffold_percentages</em>: distribution (%) of the first 50 most populated Murcko scaffolds for MBC2016, MBC2022 and ECBL. <em>Murcko_scaffolds_comparison</em>: distribution (count) of the first 94 common Murcko scaffolds for MBC2016, MBC2022 and ECBL. QikProp properties for all the analyzed DBs (6 files; CSV format). SMILES codes for all the analyzed DBs (6 files; SMI format). <em>joinplots.py</em>: python script to generate the 2D plots in <strong>Figure 2</strong> of the MS. <em>fingerprint_similarity.py</em>: python script to run and generate the Tanimoto similarity plots in <strong>Figure 3</strong> of the MS. <em>calc_kde.py</em>: python script to run kernel density analysis reported in <strong>Figure 5 </strong>of the MS.

更新说明。本次修订新增内容：用于处理数据库并计算通过Veber过滤器（Veber filter）与Ghose过滤器（Ghose filter）的分子占比的Python脚本——Veber_filter.py与Ghose_filter.py（相关数据见于手稿的表1）。本数据集包含可复现题为《MBC与ECBL文库：杰出的药物发现工具》的手稿中所有图表的数据与脚本。

待分析数据库清单：
MBC2016（总条目：1096个化合物；性质分析中剔除7.39%——因QikProp运行失败）；
MBC2022（总条目：2577个化合物；性质分析中剔除3.14%——因QikProp运行失败）；
ECBL（总条目：101021个化合物；性质分析中剔除0.20%——因QikProp运行失败）；
ChEMBL v.31（总条目：1908325个化合物；性质分析中剔除2.97%——因QikProp运行失败）；
DrugBank v.5.0（总条目：10981个化合物；性质分析中剔除4.13%——因QikProp运行失败）；
ZINC20（总条目：10723360个化合物；性质分析中剔除0.61%——因QikProp运行失败）。

相关文件清单：
1. QikProp_properties.docx：包含所有待分析数据库的QikProp性质计算完整列表的文档文件；
2. DATA_comparison.xlsx：包含可复现手稿图4所需绘图数据的Excel文件；
3. Murcko_scaffold_percentages：MBC2016、MBC2022与ECBL的前50个最富集默柯支架（Murcko scaffold）的占比（%）分布数据；
4. Murcko_scaffolds_comparison：MBC2016、MBC2022与ECBL的前94个共有默柯支架的数量分布数据；
5. 适配所有待分析数据库的QikProp性质文件（共6个，格式为CSV）；
6. 适配所有待分析数据库的SMILES编码文件（共6个，格式为SMI）；
7. joinplots.py：用于复现手稿图2的2D绘图Python脚本；
8. fingerprint_similarity.py：用于运行并生成手稿图3的谷本相似度（Tanimoto similarity）绘图的Python脚本；
9. calc_kde.py：用于执行手稿图5所报道的核密度分析的Python脚本。