Prospective Prediction of M. tuberculosis Inhibition – 3 years on

Following on from the recent blog and poster describing use of TB Mobile to predict the targets for some recently published GSK compounds, I have now looked at compounds that overlapped with some previous predictions performed 3 years ago and deposited in CDD public. In total I found 15 compounds that had antimalarial activity out of the 166 M.tuberculosis leads deposited in ChEMBL-NTD. Out of these, 2 seperate Bayesian models each predicted 11 molecules correctly. I am happy to get 73% correct although access to a larger number would have been better. I have put the quick analysis and description of using CDD to enable dataset comparison on Figshare. The models available in 2010, could have been used to screen fewer compounds. We have developed these models further in the last year so updates to come soon.

Leave a Reply

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>