Indra is one of those projects that feels like it’s going to be part of a pretty crazy future. The thing about biology is that the literature is absolutely vas

Indra is one of those projects that feels like it’s going to be part of a pretty crazy future. The thing about biology is that the literature is absolutely vast. Like millions of publications. Each paper describes some tiny piece of a cellular biochemistry using a narrative. Like protein A activates expression of gene B which then binds to and chemically modifies gene C. My personal opinion is that this mode of research is basically incapable of describing the way a cell works. It’s also strictly necessary at the moment because we’re still working out how the individual genes and proteins work. So we have a million papers that each use somewhat vague terms to term a story about the nice deterministic process by which a couple genes interact. Indra is a framework for automatically converting that natural language into precise mathematical descriptions of the biological processes. This is absolutely crucial for understanding cellular behavior because feedback and redundancy are everywhere in cellular processes. That leads to very unintuitive when one piece of the process is removed. And, removing one piece of the cellular machinery is exactly what most of our new medicines are designed to do. In the case of cancer, it’s inevitable that the cancer will exploit the feedback and redundancy of cellular machinery, finding another way to continue activating the growth pathways. I also suspect that many drug side effects are the result of changes in one part of the cellular interaction network resonating in an unexpected way far downstream from the original change. So indra, or some more advanced version of it, could pull together the biological literature into a massive computational model that allows researchers to understand how targeting one gene affects all the others.