When we founded Red Cell Partners, an investment/incubation firm focused on defense and health care, the last thing I expected to be doing was publishing oncology research. It has been fifteen years since my last academic publication, spending that time at the Department of Defense working incredibly hard problems from preventing PTSD to next-generation weapons systems. I had expected to spend most of my time evaluating and supporting early-stage startup investments. Yet, when the idea of incubating an effort to apply machine learning to cancer drug discovery arose, it was impossible not to get involved.
It haunts me that, despite exceptional efforts by an army of the most dedicated and brilliant researchers, someone dies of cancer about every three seconds and I distinctly recall the feeling of working in the field a few decades ago, where so many questions about –omics and gene functions were literally unanswerable. The human genome project, the advent of RNAi and CRISPR, the building and sharing of advanced databases, and the artificial intelligence revolution of recent years have created an opportunity to truly change the game. Re-entering the field felt like walking back into a pitch-black room with the lights now on. There is so much data out there waiting to be exploited and it is matched by an explosion of effective techniques for extracting hidden insights from that data. It truly feels like this moment may be a corollary to the discovery of antibiotics and we may be approaching our collective dream scenario: a world where a cancer diagnosis is no scarier than bronchitis or the flu.
I've been lucky to work with a team that is second to none and the paper we recently published stands for me as a proof of principle for the potential of applying machine learning to oncology, but there is so much more to be done. I love the idea of using advanced algorithms to find novel approaches to attack tumors, and I am enamored with the low cost, broad availability, and side effect profile of azithromycin. But now that the analytical architecture is built, results of that type are actually easy to find and that work is already dwarfed by the advancements we’ve made in recent months to uncover novel vulnerabilities, effective biomarkers for efficacy, and highly effective drug combinations. I could not be more excited about the potential.
On the occasions when I have a moment to reflect, I find myself oscillating between ruminations on those lost to cancer and being inspired by the incredible human effort that created the body of cancer and biotech research we all depend on today. I love the idea that cancer may become inherently self-limiting: so many of us have been inspired by a personal loss to dedicate our time and our treasure in a battle to save future generations from the pain we have suffered. I believe this may be the moment where we transition from a cancer fatality every three seconds to one every four seconds... and then five... and then six. The opportunity to contribute to that moment is the honor of a lifetime.