Project Lead(s): Izhar Wallach
There are over 6,000 neglected diseases (ND) without adequate treatments, many of which are endemic to the developing world.
The incentive to develop drugs to treat these diseases is limited because the target populations cannot afford the cost of treatment.
One possible approach to these and other diseases is to use a computer model to identify new indications for existing drugs, as this can significantly reduce development time and cost.
The first stage of this project – to identify new indications for existing drugs – involved expanding an existing computer model from a proofof-concept statistical model into a fully-fledged model.
The descriptive power of the model was elaborated to allow better representation of the features characterizing the interactions between drugs and disease targets.
The model was then to analyze and learn from a dataset 10 times larger than an earlier one used for retrospective tests.
The technology has been used in a drug repurposing setting to find new indications for approximately 14,000 approved drugs and compounds from the NCGC Pharmaceutical Collection.
The model has also been used to make predictions against the malaria targets for which structures and selected high-probability binder molecules exist.
As the process of developing the core platform technology was more extensive than originally anticipated, laboratory confirmation of the predictions has not yet occurred.
The next steps for the project are continued improvement of the predictive accuracy, and broad validation of the prediction technology. The researchers are working with a major pharma company to assess the ability of the software to recapitulate their experimental assay results and its applicability to their research pipeline.
The project has received $1,000,000 in funding from an unknown source.