admet-predictor

Predict ADMET liabilities by processing project CSV files containing physicochemical columns like molecular_weight, alogp, hba, hbd, and psa. The tool applies rule-based tiering and structural alerts to generate go/no-go classifications for candidate compounds.

Screen natural products for oral mucositis by evaluating phytochemical absorption and delivery feasibility. The method avoids blindly rejecting plant-derived compounds with poor oral Ro5 metrics by emphasizing topical or mucosal delivery route alternatives.

CYP450 interaction risk framing flags plausible drug-drug interactions by cross-referencing the project's mechanism and target CSVs against typical chemotherapy regimens. This identifies candidate compounds with potential CYP liabilities in cancer patient populations.

Assess oral versus topical delivery feasibility by analyzing PSA, LogP, and mucoadhesion properties within the inflamed-tissue context. The evaluation checks physicochemical thresholds to determine if a compound suits oral-like or mucosal delivery routes.

The toxicity screening uses an A/B/C/D/F classification system to advance or reject candidates based on rule-of-thumb thresholds and structural alert logic. This tiering framework directly supports go/no-go decisions for candidate compounds.

Avoid relying solely on rule-of-five violations for natural products with poor oral Ro5 metrics when alternative delivery routes exist. The approach emphasizes delivery-route alternatives for oral mucositis applications rather than rejecting compounds outright.