ADME

Combinatorial Chemistry and High Throughput Screening techniques allow to generate a large number of new potential drug candidates. The majority of these molecules are intended for oral therapy. Thus, there is a need to incorporate, as early as in the first phases of the drug discovery, procedures able to predict oral absorption and bioavailability of a given molecule.

Data mining models developed by BioChemics Consulting, based on Fuzzy Logic, are able to predict correctly 70-75% of the test set compounds included in oral absorption and bioavailability data sets. We hence claim ca.10-15% improvement about in the obtained validation results is, on the same data set, as compared to other prediction methods from the literature.
We have also demonstrated the importance to work with data sets with a large molecular diversity, and use appropriate prediction tools. The prediction power can be is increased by up to 25% when employing a data set with a better-optimized molecular diversity.