Transformer-CNN: Swiss knife for QSAR modeling and interpretation

Table 1 Descriptions of datasets used in the work

Code	Description	Size	Code	Description	Size
Regression tasks			Classification tasks
MP	Melting point [45]	19,104	HIV	Inhibition of HIV replication [46]	41,127
BP	Boiling point [47]	11,893	AMES	Mutagenicity [48]	6542
BCF	Bioconcentration factor [47]	378	BACE	Human β-secretase 1 (BACE-1) inhibitors [46]	1513
FreeSolv	Free solvation energy [46]	642	Clintox	Clinical trial toxicity [46]	1478
LogS	Solubility [49]	1311	Tox21	In-vitro toxicity [46]	7831
Lipo	Lipophilicity [50]	4200	BBBP	Blood–brain barrier [46]	2,039
BACE	IC50 of human β-secretase 1 (BACE-1) inhibitors [46]	1513	JAK3	Janus kinase 3 inhibitor [51]	886
DHFR	Dihydrofolate reductase inhibition [52]	739	BioDeg	Biodegradability [53]	1737
LEL	Lowest effect level [54]	483	RP AR	Endocrine disruptors [55]	930

ISSN: 1758-2946