Skip to main content

Table 6 Test sets where big property changes (logD change is above 1; solubility and clearance change is either low\(\rightarrow\)high or high\(\rightarrow\)low) are desired

From: Transformer-based molecular optimization beyond matched molecular pairs

Test set

Size

Percentage (%)

MMP

6,180

3.7

Similarity (\(\ge\)0.5)

18,546

3.9

Similarity ([0.5, 0.7))

15,130

4.6

Similarity (\(\ge\)0.7)

3,416

2.3

Scaffold

6,252

3.1

Scaffold generic

10,514

3.6

Merged

21,652

-

  1. Size indicates the number of data points where big property change are desired; Percentage indicates the fraction of the original test set in Table 2 with data points that have big property changes, e.g. 6180/166582\(\approx\)3.7%