55+ Regulatory-Grade
ADMET Predictions

Predicts mutagenic potential via bacterial reverse mutation assay in Salmonella typhimurium and E. coli strains. A core regulatory genotoxicity test required in early non-clinical safety studies.

Detects micronuclei in the cytoplasm of cells — fragments produced by DNA breakage (clastogens) or whole chromosomes displaced by disruption of the mitotic apparatus (aneugens).

Predicts clastogenic and aneugenic effects in mouse bone marrow and/or blood erythrocytes — the in vivo complement to the in vitro assay.

Predicts structural chromosomal aberrations in mammalian cells — breaks, gaps, and rearrangements observed in metaphase cells. A standard component of the ICH core genotoxicity battery.

Predicts activation of the tumor suppressor p53, a key stress-responsive indicator of DNA damage and potential genotoxicity. Widely used in regulatory screening and mechanistic hazard assessment.

Predicts stabilisation of ATAD5, which accumulates in response to DNA damage and replication stress. Provides a p53-independent readout of genotoxic potential, covering compounds that evade p53-based detection.

Rodent carcinogenicity prediction based on two-year bioassay data. Separate models for rat and mouse, both sexes, plus a general cross-species model.

Three models covering the major cardiac ion channels. NaV1.5 governs rapid depolarization, CaV1.2 contributes to the plateau phase, and hERG mediates repolarization. Inhibition may lead to QT prolongation and proarrhythmic effects.

Drug-Induced Liver Injury prediction based on human hepatotoxicity data annotated from FDA drug labels, covering serum enzyme elevations, jaundice, and hepatic failure.

Predicts inhibition of AChE, leading to acetylcholine accumulation and disruption of normal neurotransmission — relevant for pesticide and drug candidates.

Predicts nervous system damage or dysfunction caused by drugs or chemicals. Neurotoxicity is a leading cause of post-market drug withdrawals; training data covers clinical cases of neurotoxic effects.

Predicts renal damage potential. Nephrotoxicity contributes to 2% of drug failures in pre-clinical studies and 19% in phase 3 — early in silico flagging saves significant time and cost.

Predicts the GHS acute toxicity category after single oral administration in rodents, supporting hazard classification and labelling under CLP and OECD TG 423.

Predicts the potential to interfere with embryo-fetal development — teratogenicity, growth retardation, and functional deficits. Supports early prioritisation ahead of costly ICH S5(R3) reproductive toxicity studies.

Predicts cytotoxic effects in human hepatocellular carcinoma cells. Cytotoxicity may underlie organ toxicity, genotoxicity artefacts, or carcinogenic responses — a key in vitro safety model.

Predicts activation of the Nrf2/antioxidant response element pathway — the primary cellular defence against electrophiles and reactive oxygen species, and an early mechanistic indicator of oxidative-stress-mediated toxicity.

Predicts disruption of the mitochondrial membrane potential, an early marker of mitochondrial dysfunction implicated in drug-induced liver injury, myopathy, and cardiotoxicity.

Predicts activation of AhR, a ligand-activated transcription factor that triggers CYP enzyme induction and downstream toxicity pathways including immunotoxicity, developmental toxicity, and cancer. Widely used in regulatory and mechanistic hazard screening.

Five models covering full-length estrogen and androgen receptor activity, the corresponding ligand-binding-domain assays that isolate direct receptor binding, and aromatase inhibition, which governs androgen-to-estrogen conversion. Core battery for endocrine disruption screening under REACH and EPA EDSP.

Thyroid peroxidase inhibition blocks thyroid hormone synthesis, affecting neurodevelopment, metabolism, and cardiac activity. PPARγ activation drives adipogenesis and metabolic disruption. Both are high-priority endocrine endpoints.

Predicts ocular irritation and corrosion potential against GHS categories, supporting animal-free classification for pharmaceuticals, cosmetics, agrochemicals, and industrial chemicals.

Predicts the potential to cause allergic contact dermatitis. Of high regulatory relevance for pharmaceuticals, cosmetics, and industrial chemicals; aligned with OECD TG 429 and the adverse outcome pathway for skin sensitization.

Predicts compound interference with G0/G1, S, G2, and M cell-cycle phases. Alterations are closely linked to genotoxicity, carcinogenicity, and cytotoxicity and are of high relevance for safety assessment.

Three models covering compound-induced effects on human hematopoietic cell lines (RPMI-8226, MOLT-4) and the pro-inflammatory cytokine IL-1B — relevant for immunosuppressive, immunostimulatory, or pro-inflammatory hazard assessment.

HBA, HBD, molecular weight, LogP, TPSA, Fsp³, rotatable bonds, aromatic rings — the full descriptor set for drug-likeness assessment.

Lipinski Rule of Five, QED score, Drug-likeness, SA Score, PAINS alerts, GSK 4/400 Rule, Pfizer 3/75 Rule — multiple filters in a single computation.

Composite score (0–1) ranking compounds by their likelihood of pharmaceutical acceptance. Derived from: Ames mutagenicity, genotoxicity (in vitro & in vivo micronucleus), carcinogenicity, hERG cardiotoxicity, DILI, DIN, AChE inhibition, estrogen & androgen receptor binding, BBB permeability, PPB, Caco-2, HIA, metabolic stability, and aqueous solubility.

Composite score (0–1) estimated from predicted mutagenicity, genotoxicity, rodent carcinogenicity, and key physicochemical properties. Lower score = lower carcinogenic risk.