alvaDesc

alvaModel is a software tool for building, validating, comparing, and applying QSAR and QSPR models. It supports regression and classification workflows based on molecular descriptors and fingerprints, with a strong focus on model transparency, interpretability, and scientific robustness. The software includes multiple data splitting strategies, variable selection methods, modeling algorithms, and comprehensive internal and external validation procedures. alvaModel provides diagnostic plots, applicability domain analysis, and model comparison tools to support the identification of reliable and predictive models. Designed according to best practices in chemometrics, alvaModel facilitates the development of interpretable models consistent with the OECD principles for QSAR validation, making it suitable for research and regulatory-oriented applications. The graphical interface guides users through the entire modeling workflow while allowing full control over each modeling step.

alvaMolecule is a no-code cheminformatics tool for visualizing, curating, and standardizing molecular datasets before analysis. It supports common molecular formats (SMILES, SDF/MOL2) and lets users explore collections in grid or spreadsheet views, with automatic import of associated data. The software provides structure verification and standardization using predefined standardizers and custom SMIRKS rules, helps detect and manage duplicates, and offers scaffold analysis to summarize core frameworks. Built-in filters and charting tools enable sorting by substructure, calculated molecular descriptors, and physicochemical properties. alvaMolecule calculates ~88 structural and physicochemical properties, including drug-like and lead-like scores such as LogP, TPSA, and the Lipinski alert index, helping prepare high-quality datasets for QSAR/QSPR modeling, descriptor calculation, and virtual screening workflows.