The SCAM Detective application provides an alternative method for assessing the potential of chemicals to be putative aggregators and cause false-positive readouts in bioassays. SCAM Detective makes predictions based on Quantitative Structure-Interference Relationship (QSIR) models build on two independent datasets generated from High Throughput Screening campaigns against AmpC β-lactamase (PubChem AID 485341/485294 and AID 585/584) and the cysteine protease cruzain (PubChem AID 1476/1478). The models were developed using open-source chemical descriptors based on ECFP6-like Morgan fingerprints with 2048 bits and an atom radius of 3 calculated in RDKit, along with the random forest (RF) algorithm, using Python 3.6. The models were generated applying the best practices for model development and validation widely accepted by the community. Batch processing is available through https://github.com/alvesvm/scam_detective. For more information, please refer to our paper: Alves, V. M.; Capuzzi, S. J.; Braga, R.; Korn, D.; Hochuli, J.; Bowler, K.; Yasgar, A.; Rai, G.; Simeonov, A.; Muratov, E. N.; Zakharov, A. V.; Tropsha, A. SCAM Detective: Accurate Predictor of Small, Colloidally-Aggregating Molecules. J. Chem. Inf. Model. 2020, acs.jcim.0c00415. https://doi.org/10.1021/acs.jcim.0c00415.

Instructions