Numerate awarded new contract by Defense Threat Reduction Agency

Sep 10 2010

Numerate, Inc. announced today that it has been awarded a new three-year contract by the Defense Threat Reduction Agency (DTRA), an agency within the U.S. Department of Defense, to adapt its pioneering small-molecule drug design platform to the rapid design and optimization of drugs against biowarfare agents.

“Drugs targeted against bio-threats need to meet a high standard in terms of safety, efficacy and dosing convenience, as well as have minimal potential for drug-drug interactions”

Numerate may receive up to $6.75 million to support its development efforts under the new DTRA contract, which is funded as part of DTRA's Transformational Medical Technologies Program (TMT). Under the contract HDTRA1-10-C-0029, Numerate will extend its breakthrough computational methods for designing active and selective lead compounds by developing the capability to simultaneously optimize absorption, distribution, metabolism, and elimination (ADME) profiles while eliminating the potential for drug-drug interaction liabilities. Numerate scientists will utilize the developed technology to deliver novel small molecule drug leads for treating the hemorrhagic fever virus Ebola, as well as an anti-bacterial drug lead with broad-spectrum activity.

Modern Machine-Learning Techniques in Drug Design

Numerate's drug design platform consists of proprietary and validated methods for using biological assay data to develop predictive models of activity with accuracies comparable to laboratory testing. These models enable the rapid design of small molecule lead compounds in silico. The result of a $30 million investment and ten years of work, the platform leverages a set of proprietary algorithms that provide accurate predictions of molecular properties. These algorithms include modern machine-learning methods similar to those used in finance, web-search, ad placement and product recommendations by companies such as Google, Amazon and Netflix. The technology resolves the issues of noise and bias in biological data, and its ability to predict the biological activity of a compound has been extensively validated in the laboratory.

The DTRA award will enable Numerate to apply its technology to develop new methods for predicting the ADME properties of compounds, enhancing the effectiveness and safety of drug leads that could lead to new biowarfare countermeasures.

"Drugs targeted against bio-threats need to meet a high standard in terms of safety, efficacy and dosing convenience, as well as have minimal potential for drug-drug interactions," said Dr. Nigel Duffy, Numerate's Chief Technology Officer. "By extending the capabilities of our technology to address ADME issues, we expect to deliver a cutting-edge drug design system that can provide unprecedented speed, efficiency and predictability in designing biowarfare countermeasures."

"We are excited to be working on this project with TMT one of the few organizations with a vision to radically transform the drug discovery and development process," said Guido Lanza, Chief Executive Officer at Numerate. "The only way to significantly improve the efficiency of drug discovery is with a new approach that lets the data drive the design process and addresses all of the key design criteria in parallel, including activity, ADME, patentability, and safety. Numerate's technology platform provides the basis for just such a paradigm shift, not only to enable the timely design of biowarfare countermeasures, but also to solve major challenges in the design and development of new drugs to treat cancer, diabetes, inflammation and other diseases burdening the healthcare system."

Leveraging the Computing Power of the Cloud

To make their discoveries, Numerate scientists begin with a quantitative set of requirements for a new drug for a given indication. Next, they obtain and curate all available data relevant to these requirements and develop predictive models for each property using modern machine-learning techniques. They then proceed to search for molecules likely to meet the desired profile by evaluating spaces of up to 10 billion compounds with these predictive models. Finally, the best compounds are synthesized and their properties verified in the laboratory.

"This process is extremely computationally intensive, requiring up to one million CPU hours per project. Such a process would be inconceivable without capabilities such as Amazon's Elastic Compute Cloud (EC2), which we have leveraged since Numerate's inception three years ago," stated Dr. Duffy.

Numerate will expand its use of the cloud to optimize the ADME properties of these compounds to ensure that their successes in vitro routinely translate to successes in vivo and in the clinic.

Source Numerate