applicability-domain

Protein target prediction using random forests and reliability-density neighbourhood analysis

A cheminformatics package to perform Applicability Domain of molecular fingerprints based in similarity calculation.

Classification models for hemolytic nature and hemolytic activity predictions in peptide/protein sequences

Reference implementation of the Distance-Based Boolean Applicability Domain for HTS datasets

Allows to visualize and analyze if the molecules of the test set and of an external set are contained in the convex hull defined by the molecules of the training set.

Reference implementation of the Vanishing Ranking Kernels (VRK) method

One fixed 9-stage QSAR and conformal-prediction drug-discovery pipeline run unchanged across ENPP1, NLRP3, TYK2 and IRAK4 — its most valuable outputs are the two refusals.

pDILI_v1 is a python package that allows users to predict the association of drug-induced liver injury of a small molecule (1 = RISKy, 0 = Non-RISKy) and also visualize the molecule.

Applicability Domain Methods of Viral Load and CD4 Lymphocytes.

This contains an end-to-end ML pipeline that predicts the activity (active vs inactive) of chemical compounds against the Human Angiotensin-Converting Enzyme, a major hypertension target.

QSAR and cheminformatics study of DPPH antioxidant activity using molecular fingerprints, physicochemical descriptors, scaffold- and DOI-aware validation, applicability-domain analysis, and sulfonamide–Tyr–Gly case studies.

Code, curated data, and frozen outputs for 'Where and when a molecular property model can be trusted': random-forest disagreement ranks activity-model error in distribution, degrades unevenly under temporal shift, and should not be used as a conservative acquisition rule. Every reported number is machine-verified.