Note: Thank you to Bill Busa, CEO and co-founder of EvE Bio, for an extremely helpful discussion while working on this essay.
This essay is long, and I recognize that many people don’t necessarily care about the details. The real headline point you need to be aware of is this dataset, which was produced by EvE Bio underneath a CC-NA license, and is a comprehensive mapping of the interactions between a significant fraction of clinically important human cellular receptors and 1,600~ FDA-approved drugs. I strongly believe that this data is really, really useful, and more people should be aware it exists.
If you’d like to understand why I think it is useful, and what the dataset exactly contains, read on!
Introduction
If you were to be a fly on the wall during the 1-6 years of preclinical drug discovery research within a pharmaceutical company, one observation you may walk away with is that, while the work is certainly complicated, it is also frighteningly limited in scope. What you’ll learn is that drugs are made by corporations that are optimizing for one primary thing, and one thing only: work. ‘Working’ is obviously contextual, but it is a simple concept no matter the situation: reduce a worrying biomarker, improve mood, lengthen lifespan and so on and so on. What does this discovery process ignore? Simply put: everything else a drug could do beyond that.
Yes, that’s a roundabout way of describing ‘off-target effects’ — defined as the action of a drug at a gene product other than the gene product it was intended to affect — but I think it’s a helpful intuition pump. Viewing the drug discovery process as ‘not paying attention to anything that is unrelated to the drug working’ is useful in that it contextualizes the situation we’re in. Drugs are meant to make money, and money is derived from drugs working. To spend time on understanding what else a particular drug does beyond It Working for its intended task is time lost and money lost.
One unfamiliar with the drug discovery process may find this bizarre; why wouldn’t the well meaning scientists in charge of developing drugs try to deeply understand how it interacts with the body? On the other hand, those deeply in the medical field would find this thesis so obvious that stating it is unnecessary; of course a pharmaceutical company would limit their scope of understanding a drug to things that lie between it working and not working. There’s only so much time and resources to go around. Priorities!
Of course, if an off-target effect comes between the drug and It Working, then certainly resources will be allocated to deal with it. But beyond that, mapping everything else a clinical-stage drug does — every receptor it unintentionally binds, every pathway it nudges sideways, every gene it perturbs slightly — is deemed so high effort and so low ROI, that it is relegated to hoping an academic will study it. Only if post-marketing surveillance turns up something worrying shall further exploration occur. Because, again, a deep understanding of what exactly an exogenous chemical is doing inside a body is not the point of the drug discovery process. Working is the point!
With that background context, I am ready to present three claims I’m going to make in this essay and spend the remaining sections trying to prove:
Understanding off-target effects is really useful. Learning about off-target effects at scale is possible. No for-profit institution has a strong incentive to do this work.
... continue reading