ConFiLiS
Consensus Fingerprints for Ligand-based Screening
Consensus fingerprint similarity screening for ligand-based virtual screening.
About
ConFiLiS v1.0 is an open-source ligand-based virtual screening tool that combines multiple molecular fingerprint types — Avalon, Circular (ECFP6) and PubChem Substructure — to identify candidates structurally similar to a query molecule.
It uses Euclidean distance with z-score normalisation across fingerprint types, and ranks candidates by a consensus average score. This consensus approach reduces fingerprint-specific biases and increases chemical diversity among the top-ranked hits.
Originally developed and applied in the dimethoxycurcumin / ABCC3 (MRP3) cancer drug-resistance study (Nelen et al., 2025), ConFiLiS is now released as a standalone reproducible tool with documented preprocessing and usage instructions.
Lead authors: Jochem Nelen
Key features
Three fingerprint types
Avalon, ECFP6 and PubChem Substructure fingerprints combined per query.
Consensus ranking
Z-score-normalised distances averaged across fingerprints to mitigate single-method bias.
Reproducible preprocessing
Charge normalisation, salt removal and tautomer standardisation built in.
Access
HPC tool
Source code on GitHub
Open source · MIT-style licence · documentation in repo
https://github.com/Jnelen/ConFiLiS →Cite this tool
Targeting Drug Resistance in Cancer: Dimethoxycurcumin as a Functional Antioxidant Targeting ABCC3