Table 3.

Statistics of the initial ChEMBL data set and filtered ones by SCScore values and the number of the resulting distinct fragment and context pairs

Data set	Number of compounds	Number of distinct fragments & contexts of radius 3
ChEMBL	1,557,992	51,730,960
SCScore <= 3.5	552,162	20,514,883
SCScore <= 3	284,461	10,661,179
SCScore <= 2.5	111,365	4,091,634
SCScore <= 2	27,916	951,993