Big Data Analytics and Machine Learning techniques are starting to achieve dramatic impact on drug discovery and precision medicine.

Pumas.ai is a bootstrapped venture at the cutting edge of the field. I learned much from this conversation. You will too!

Sramana Mitra: Let’s start with an introduction of yourself as well as the company.

Joga Gobburu: I’m a faculty at the University of Maryland. Before that, I used to work at the FDA. At the university, I came up with inventions. That led to an entrepreneurial spirit.

Sramana Mitra: What was the invention?

Joga Gobburu: The core focus is on predictive healthcare analytics. The tools that we have are clearly outdated. They are not attuned to the demands of pharmaceutical Drug Development and patient care today. Naturally, we saw an opportunity. Also, we came up with innovations that are enabling scientists to gain deeper insights for developing drugs much faster and to enhance precision medicine to patients.

Sramana Mitra: Are you applying this innovation to drug development primarily?

Joga Gobburu: Yes, primarily in drug development and patient care in the hospital setting.

Sramana Mitra: Let’s do the drug development use case first. What is the dataset?

Joga Gobburu: I need to take a step back and explain a little bit more about the data analytics tool that we have developed. We have a tech partner which is a spin off from MIT. Pumas was developed in Julia. We have a very strong connection with MIT as well as MIT spin offs.

Coming back to your question about the data, we have three offerings. One is called Pumas. Pumas.ai is the name of the company and it has three offerings. Pumas is the pharmaceutical modeling and simulation software. It’s a suite of tools. The second offering is Lyv which is the clinical decision support system for personalized medicine. The third one is our consulting services. We work with pharmaceutical companies and other research organizations to develop drugs.

In terms of the application of Pumas, they are used through the drug development process from discovery to lifecycle management. As you gather information from early discovery about a new molecule or a new set of molecules, it takes about 10 to 12 years for a new treatment to be approved for marketing. It costs over a billion dollars.

Naturally, one of the important maneuvers is derisking. You leverage prior knowledge. What’s known about the type of molecule or similar molecules? Maybe somebody else has done a similar experiment. Then you leverage that knowledge to make narrow down your choices from infinite choices. That’s where Pumas comes in.

Pumas allows scientists to take this heterogeneous dataset and extract signals from that to project the likely outcomes and allow them to move forward in drug development. The data comes in different forms depending on the stage of drug development. Data comes in as lab parameters, lab testing results, and molecular structure.

As the molecule progresses through the drug development, you get information from animals, big animals, and then humans. How does the body react to a new molecule? How does the patient feel after they take the medicine? How do you take all this information and use that to come up with the best way to test new treatments?