From Commercial to R&D and everything in between — my journey in GSK’s Future Leaders Programme
Originally published on the GSK Tech Blog, October 30,2019
When I was studying biomedical engineering in undergrad, the pharmaceutical industry wasn’t always on my radar. I thought it wasn’t for me — while I loved learning about science and knew that these companies were doing amazing research, I didn’t love being in a lab all day, and I didn’t exactly want to go to graduate school, either.
That all changed when I interned at a pharmaceutical company during the summer before my final year. I had accepted the position because it sounded like I would be doing some interesting analysis on clinical trials. That summer, I discovered a part of the industry that I had never seen in class — the applications of technology in pharma R&D. My colleagues were using everything from IBM Watson for drug discovery to AI and machine learning in imaging and human genetics. I was fascinated — what was this whole new world where I could combine my love of science with technology to help patients? My internship experience, with the support of my colleagues there, led me to where I am now at GSK, as an associate in the Tech Future Leaders Programme.
The Tech Future Leaders Programme (FLP) is a development programme that allows recent university graduates to rotate through GSK Tech departments to gain a breadth of knowledge on the applications of technology across different business units. As FLPs, we are provided with a support network and mentors to help us guide our careers, and a certain degree of fluidity for our rotations. Thanks to the programme, I have had the unique opportunity to explore both the commercial and the R&D sides of GSK.
For my first rotation, I was a web developer and worked with external agencies and our marketing teams to create websites for healthcare providers. It was a great learning experience to understand the commercial operations of a business and the nuances of a large organization. I joined Tech during a period of transformation — our Chief Digital and Technology Officer, Karenann Terrell, has been implementing changes to make GSK more technologically and digitally advanced and agile. To me, this transformation has been exciting — new ideas, products, teams, and leaders have brought some challenges, yet so many meaningful projects have emerged. With everything from our partnership with 23andMe to creating machine learning solutions to improve our supply chain, it has been thrilling and gratifying to see our tech projects have a direct impact on how we get information and medicines to patients.
When it came time to find my next rotation, I found myself yearning to go back to R&D (I am still a science nerd at heart!), and to explore that interest that developed during my internship. Surely there were teams doing similar things here at GSK, maybe even more! I perused org charts, sent emails and made introductions to colleagues in the data space. I was unsure of how they would be received, but more than willing to put myself out there and take that chance.
I have been so pleasantly surprised by the positive reception! Colleagues have been eager to get to know me, hear my interests, and direct me to others who could give more clarity on my career path. This honesty, enthusiasm, and humanity speak to the culture GSK fosters and the wonderful people who have led me to where I am today. A casual meetup I had with a Tech colleague that was supposed to be about opportunities in data engineering resulted in talking about her secondment at this R&D group called ATOM.
I was so intrigued, and after learning more about it, I knew it was something I wanted to be a part of. After much conversation with FLP and ATOM colleagues, I received full support to join the ATOM team in San Francisco for my second rotation.
ATOM — What we’re doing to accelerate drug discovery
The Accelerating Therapeutics for Opportunities in Medicine (ATOM) Consortium is a public-private partnership between GSK, Lawrence Livermore National Laboratory (LLNL), Frederick National Laboratory for Cancer Research, and University of California, San Francisco. We have an ambitious goal of accelerating drug discovery from a slow, sequential, and high-failure process to an integrated approach that can one day deliver a drug to a patient in less than a year using high-performance computing and machine learning models. GSK has donated data from over 2 million compounds in its historical and current screening collection, and that data, combined with public datasets and data from future consortium members, is the foundation for the tools we build.
One of the projects I’m involved with has a goal of optimizing preclinical safety predictions so we can incorporate predictive toxicology and computational modelling early in the drug discovery process. Drug induced liver injury (DILI), is a leading cause for attrition during drug development and one of the reasons why drugs are withdrawn from the market. DILI is the most common cause of death from acute liver failure and accounts for approximately 13% of cases of acute liver failure in the United States. There are currently no reliable methods to screen potential drug candidates for DILI preclinically, and the specifications used during high-throughput screening are often limited in scope. Using historic toxicology data from GSK, ATOM-generated data, and public datasets, we are incorporating quantitative systems toxicology tools and machine learning to predict DILI from the structure of a proposed drug lead.
The DILI project is just one of many we’re trying to tackle at ATOM. Another sub-team is creating generative molecular design loops, which involves a scalable, active learning workflow to train and retrain machine learning models (see figure above!). With this platform, we have been able to screen over 3 million compounds in under 24 hours, using just 6 nodes in a LLNL supercomputing cluster. The LLNL supercomputing clusters are some of the world’s most powerful. Once the platform is scaled up, we will have the capacity to screen billions of molecules at a much faster rate than before!
It personally feels awesome to be back in the depths of science and be learning so much every day. It’s crazy to think that 3 years ago, I wasn’t even considering pharma to be a potential career path — and now I’m working alongside some incredible scientists to change the landscape of drug discovery. It’s an honour and a privilege to know that my work is directly impacting patients, and I’m thankful to have such a supportive community around me to keep pushing me to learn and be my best.
If you’re interested in learning more about ATOM, check out our website, LinkedIn, or Twitter.
If you’re interested in learning more about the Tech Future Leaders Programme, please visit GSK Careers.
The ATOM preclinical drug discovery workflow
Special thanks to the GSK Tech blog team, Antonio Rocca, Dave Brown, Shalin Simmons, John Katsnelson, Andrew Weber, and Stacie Calad-Thomson for feedback and advice on this blog.
Jillian Yong
Data Scientist, Tech Future Leaders Programme, GSK
