Validating New Tuberculosis Computational Models with Public Whole Cell Screening Aerobic Activity Datasets

  • Sean Ekins, Joel S. Freundlich
  • Pharmaceutical Research, March 2011, Springer Science + Business Media
  • DOI: 10.1007/s11095-011-0413-x

TB whole cell models for aerobic activity

What is it about?

We used a small TB whole cell data set published by Novartis (aerobic and anaerobic hits) as a test set for previously published Bayesian models. We demonstrated enrichment over random screening. We also used a published dataset of FDA approved drugs as a test set.

Why is it important?

This represents one of the early validation studies for the bayesian models which we built with public TB datasets. We also built a model with this small dataset to identify aerobic hits.


Dr Sean Ekins
Collaborations in Chemistry

Some of the additional data in this paper is of interest such as the analysis of the FDA drugs based on lipinski violations as well as SMARTS filters analysis of different TB datasets. This represented one of our early comprehensive efforts at model validation with external test sets published by other groups.

Read Publication

The following have contributed to this page: Dr Sean Ekins and Joel Freundlich