Watson for Drug Discovery
What is Watson for Drug Discovery?
Watson for Drug Discovery uses machine learning to find relationships between genes, drugs and other entities. Since Watson is able to ingest massive amounts of data points, it’s able to uncover relationships between the entities more efficiently than humans. This doesn’t replace the expertise of a drug researcher, but aides him in the discovery process.
See the Forbes article.
See the product page.
My role and business objective
The process of drug discovery can take years. Lab and literature based research must be meticulously submitted to the FDA. Many iterations can take place between the drug company and the FDA before clinical trials begin.
The researchers at Pfizer needed a way to shorten their research time. Watson gave them the ability to search many resources at once.
The key users and how they were identified
An advantage to working on client contracts is that the contract almost always comes with an assigned set of users from the customer. Pfizer was no different.
Two key sponsor users were Maryann Whitely, a researcher and Dhiraj Gambhire, an MD who helped run the clinical trials.
How I practiced Design Thinking
Luckily, for this project I was paired with Will Walkington, a designer in Austin who works on the Watson for Drug Discovery product. We were building the Pfizer solution using WDD as it’s base. So Will was a great asset.
Will and I both traveled to Cambridge to conduct interviews. We then held a design thinking workshop in San Diego with a team of scientists, researchers and doctors. From IBM we had two scientist SMEs, the offering manager, NLP (natural language processing) experts, back end developers and managers.
The actions I took
Using an agile process, Will and I were able to each do two sections of the app. One around efficacy and one around toxicity. The customer was involved in design reviews from the conception of wireframes through the high fidelity solutions that were delivered to the development team.
Once the wires were approved, I created the final high fidelity assets and delivered to development. The process was done in an agile method and tracked with Epics, stories and tasks.