Previous Article in event
Next Article in event
Next Article in congress
Predictive models as a useful tool for preclinical assay optimization in antimalarial compounds.
Published:
21 October 2021
by MDPI
in MOL2NET'21, Conference on Molecular, Biomed., Comput. & Network Science and Engineering, 7th ed.
congress CHEMBIO.INFO-07: Cheminfo., Chemom., Comput. Chem. & Bioinfo. Congress München, GR-Cambridge, UK-Ch. Hill, USA, 2021.
Abstract:
In this study, three Perturbation Theory Machine Learning (PTML) models were created to optimize preclinical assays on antimalarial compounds of the parasitic species of the genus Plasmodium falciparum. Between General Discriminant Analysis (GDA), Classification Tree with Univariate Splits (CTUS) and Classification Tree with Linear Combinations (CTLC). The PTML-CTLC presented the best performance with a Sensitivity percentage equal to 83.6 for the training data set and 85.1 for validation; for specificity with a percentage of 89.8 for training and 89.7 for validation. The PTML-CTLC model has significant variables that could be a good option for pharmaceutical companies to optimize preclinical testing processes.
Keywords: Perturbation Theory, Machine Learning, PTML, Plasmodium falciparum
