Joshua Welch has cracked open a problem that has long stumped cellular biologists: figuring out which cell is talking to which. Working at the University of Michigan, Welch and his team have developed CytoSignal, an algorithm that does something previously thought nearly impossible—it reveals cell-to-cell communication at the level of individual cells, watching those conversations unfold across time. The research, published in Nature Genetics, marks a fundamental shift in how scientists can observe the silent but critical conversations happening inside living tissue.
Cells communicate constantly, exchanging information to coordinate growth, determine identity, and respond to their neighbors. Until now, scientists could detect these conversations only in broad strokes—looking at interactions between groups of cells, or freezing their observations at a single moment in time. It was like trying to understand a conversation without being able to hear the actual words or watch it unfold. CytoSignal changes that. By analyzing spatial transcriptomics—maps showing where genes are expressed throughout tissue—the algorithm can identify which specific cell is signaling to which neighbor, and track how those signals change over hours and days.
What makes CytoSignal particularly powerful is its ability to distinguish between two fundamentally different ways cells communicate. Some cells must physically touch to exchange information—contact-dependent signaling. Others send out protein messengers that diffuse through space like chemical whispers. CytoSignal can tell them apart. The algorithm works by identifying ligand-receptor pairs: ligands are the molecular "keys," and receptors are the "locks" on cell surfaces. When they fit together, communication happens.
To test their tool, Welch's team studied embryonic mouse tissue, a system where cell communication is absolutely critical. The cells in a developing embryo must coordinate exquisitely, with each cell receiving cues about its role and position. CytoSignal predicted which ligand and receptor proteins would bind at specific locations in the tissue. The researchers then used antibodies that fluoresce when they find actual ligand-receptor binding, creating a glowing map they could overlay on CytoSignal's predictions. The match was striking—the algorithm's predictions aligned with biological reality.
The implications ripple outward immediately. Many existing drugs work by targeting receptors on cell surfaces. If researchers can use CytoSignal to identify which receptor signaling is broken in disease states—comparing healthy tissue to diseased tissue—they gain a direct pathway to new treatments. Welch himself sketches a more ambitious future: "The technique can be used to understand how genetic mutations or disease states disrupt normal signaling. Down the line, you could even use CytoSignal to develop cell therapies by turning off defective signaling."
Perhaps most fundamentally, CytoSignal liberates scientists from having to guess. Earlier approaches required researchers to pick a single ligand-receptor pair they suspected was important, then study it in isolation. This new method requires no such hunch. Welch explains the shift: "You don't have to go into an experiment knowing which ligand is involved—you can look at all of the genes and then determine from the data which are most important." The algorithm, in other words, lets the tissue itself reveal its secrets. As spatial transcriptomic data becomes cheaper and more abundant, CytoSignal offers a tool to extract meaning from it—to listen, finally, to what cells are actually saying to one another.