Table 1.

Examples of formalized MCO applications in drug discovery that maximally exploit different features of the desirability principle

Application	Endpoints being co-optimized	Desirability function feature(s) that is (are) most exploited	Refs
Central nervous system marketed drugs	Lipophilicity; distribution coefficient; topological polar surface area; molecular weight; number of hydrogen bond donors; most basic center	Solution ranking and VS	[33]
Non-nucleoside HIV reverse transcriptase (RT) inhibitors	RT inhibitory efficacy; toxicity over MT4 blood cells	Solution ranking and VS	[54]
Antidepressant drugs	Binding to a targeted receptor (tR); functional assay on a receptor different to tR; binding to other four receptors; probability of non-mutagenicity; metabolization rate	Adaptability	[28]
Inflammatory/immune process (P2X7 inhibitors)	Potency; solubility; safety	Ability to deal with missing values and data uncertainty; avoiding hard filters	[26]
Antibacterial activity (fluoroquinolones)	Potency; safety; bioavailability	Solution ranking and VS	[35]
Central nervous system marketed drugs	Aqueous solubility; human intestinal absorption; calculated logP; P-gp transport; plasma protein binding; CYP2D6 affinity; CYP2C9 affinity; blood–brain barrier penetration; hERG inhibition	Ability to deal with missing values and data uncertainty; avoiding hard filters	[29]
Antialzheimer agents	Affinity; selectivity	Solution ranking and VS	[55]
Extended-release formulations for propranolol	Set of 20 pharmacokinetics parameters	Adaptability; avoiding hard filters; solution ranking and VS	[27]
Optimization of oral drugs	MOLECULAR weight; ALOGP; number of HBDs; number of HBAs; molecular PSA; number of ROTBs; number of AROMs	Solution ranking and VS	[56]
Inhibitors of serotonin 5-hydroxytryptamine (5-HT1A) receptor	Set of 11 pharmacokinetics parameters	Ability to deal with missing values and data uncertainty	[57]