Leveraging heterogeneous data from GHS toxicity annotations, molecular and protein target descriptors and Tox21 assay readouts to predict and rationalise acute toxicity

Table 1 Top 5 enriched “reactive, unstable, toxic” endpoint ToxAlerts substructures in the binarized toxic set (GHS acute oral toxicity category of 1–3) versus nontoxic set (acute oral toxicity category of 5 or implied nontoxic)

Alert ID	Description	Odds ratio	p value	Source
TA1000	Double P=S and P=C bonds	35.2	4.7 × 10⁻¹⁹	Enamine
TA975	Thiocarbonyls	27.6	9.6 × 10⁻⁶	Enamine
TA880	Gem-Dihalo propane and cyclopropane	25.1	3.1 × 10⁻⁵	Life chemicals
TA567	Thioureas	25.1	3.1 × 10⁻⁵	ChemDiv
TA885				Life chemicals
TA1075				Ontario institute for cancer research
TA914	Nitrosos	16.4	5.0 × 10⁻⁶	Life chemicals
TA998				Enamine
TA1089				Maybridge

Only enrichments with a p value below the Bonferroni-corrected cut-off equivalent to α = 0.05 were considered. The remaining significant enrichments were ranked according to their odds-ratio, or effect size. These alerts generally represent reactive functionalities that might be anticipated to afford nonspecific toxicity

ISSN: 1758-2946