Our research activity

PIKAÏROS conducts applied research in cheminformatics and computational drug discovery. Our work focuses on the development and validation of predictive models for ADMET properties, virtual screening methodologies, and machine learning approaches to chemical data analysis.

Research areas

Virtual screening and hit identification

Development of ligand-based and structure-based scoring approaches for early-stage drug discovery. This includes scaffold-based chemical space exploration, novelty benchmarking against curated reference sets, and multi-criteria candidate prioritization.

ADMET modelling

Development of QSAR and QSPR models for the prediction of absorption, distribution, metabolism, excretion, and toxicity properties from molecular structure and physicochemical descriptors. Work includes handling of imbalanced datasets, applicability domain assessment, and different strategies for model validation.

Machine learning for chemical and biological data

Application of supervised and ensemble machine learning methods to cheminformatics problems, including classification on highly imbalanced HTS data, consensus modelling strategies, and explainability approaches relevant to regulatory and scientific scrutiny.

Bioinformatics and data processing

Computational analysis of biological datasets including mass spectrometry (differential protein expression), sequencing signal processing, and microscopy image analysis.

Our scientific papers

Reproducible KNIME workflows associated with published studies are available for download on the individual publication pages. These cover ADMET prediction, solubility modelling, and QSAR model construction for imbalanced datasets.

See all our papers