Main criteria	Description	Score
Data quality	Relevance: 0, 1 for borderline, 2 for exact	0-3
	Quality: 0 for no-info, 1 for good quality (applied only if relevance >0)
Chemical number	0: <100,1: 100-500, 2: 500-5000, 3: >5000	0-3
Descriptors/fragments	0: no info	0-3
	1: only partial info available
	2: possible ambiguities depending on the chemical format
	3: full description, equation available
Explicit and verified the algorithm	0: no info	0-3
	1: only partial info available
	2: possible ambiguities
	3: full description, equation available
Applicability domain	0: no info	0-3
	1: only partial info available
	2: explained, but to be applied manually
	3: full description, and model provide tool
Performance	0: R2<0.65,1: 0.65<R2<0.85, 2: R2>0.85	0-3
	+1 added if the training set is available
Validation	0: Q2<0.60,1: 0.60<Q2<0.80, 2: Q2>0.80	0-3
	+1 added if the external validation set is available
Output	+1 if univocous	0-3
	+1 if usable as it is
	+1 if usable as key study or not
Cost	0: annual license and cost, 2: perpetual license, 3: free	0-3