A Rhodes scholar from Stanford’s physics and computer science departments, Emma uses computational statistics to study complex problems. One of the most difficult is cancer. Emma explains that the complexity of the disease involves detailed descriptions of pathways, each path a tree of maybe 30 different molecules, and each branch of the tree labeled as activating or inhibiting cancer. “The amount of knowledge obtained is far beyond what anyone can grasp in a thousand lifetimes,” she says, “and also impossibly superficial.” She believes that the edges of the trees ought to be labeled not with a binary variable, but with a function that depends on hundreds of parameters—the cellular environment, the concentrations of reactants, and the type of cells.
To that end, Emma is re-figuring the trees as graphs. She thinks of each tree not as independent, but entwined in a forest. The pathways intersect and influence each other. “So what we really have,” she says, “is a gigantic graph with thousands or millions of molecules, and each edge with thousands or millions of parameters.”
“Every tumor represents a slightly different failing in a diabolical spider web of parameters all called cancer. There isn’t a chance we’re going to make sense of it all without using a computer.”