Photoreceptor cells. (Image: Mesa Schumacher)

Biologists have long sought to understand human cells, and in recent years, tools like high-throughput genetic sequencing have afforded unprecedented access to the underlying molecular machinery. Those efforts have produced a torrent of data in a brief amount of time. In response to that surge in cell sequencing, researchers have launched large-scale efforts to collect those data including the Human Genome Project, the Human Cell Atlas, and the Human Protein Atlas.

Those projects need something besides terabytes: A fundamental framework that can connect them all.

“Without a proper theoretical understanding, they are confronting this deluge of data without a meaningful way to organize it,” says biologist and SFI External Professor Manfred Laubichler (Arizona State University).

Two years ago, Laubichler helped organize an interdisciplinary SFI working group that over the course of a few days developed a radical new organizational structure for cell types. Instead of categorizing cells by phenotypes, or observable features, the researchers grouped cell types along new conceptual lines. They proposed grouping cells together according to their shared evolutionary history, and by identifying the molecular agents that determine the ultimate fate of a cell.

The meeting led to a pivotal paper published in Nature Reviews Genetics in 2016. Now, it’s inspired a second working group, held in March at SFI, designed to put those conceptual ideas to work. Participants will include the biologists and physicists who helped develop the original idea, as well as data-driven scientists engaged in big data efforts like the Human Cell Atlas.

“We’re bringing new people into the SFI orbit,” says Laubichler.

The working group has two main goals.

First, they want to hammer out a better understanding of the formal framework proposed in the NRG paper. Second, they want to produce a functional model that embodies the proposed new concept of cell types. Data scientists might use such a model to analyze sequencing data to identify patterns and connections among cell types based not on appearance, but instead on shared evolutionary history.

Such a model, says Laubichler, would bring organization and structure to those ongoing projects. “We want to operationalize the concept so it becomes useful for all those large-scale efforts,” he says. “This is quite an urgent and important thing to do at this stage.”

Read more about the Working Group "Operationalizing the Cell Type Concept."