The bodies of most animals contain repeated segments, such as vertebrae or ribs. As the embryo develops, these units are laid out sequentially, from head to tail with impeccable timing.

Scientists have long known that expression goes up and down for many genes in cells of the pre-somitic mesoderm — the tissue in the vertebrate embryo from which segments arise. This oscillation in gene expression parallels the rhythm of segment formation, as if each cell were keeping time with an internal clock. But although much is known about the molecules that regulate oscillation inside each cell, scientists have not yet worked out how cells synchronize their clocks so that they all tick in unison.

Kazuki Horikawa took a two-pronged approach to this question. He and colleagues at the University of Tokyo in Japan put genetically engineered cells from the pre-somitic mesoderm into zebrafish embryos to monitor the effects of different molecular changes on segment formation and gene expression. At the same time, collaborators at Nagoya University constructed a mathematical model of the multicellular clock to predict how changes in one parameter would affect clock dynamics. As described on page 719, the teams took turns conducting virtual simulations and in vivo experiments, each process informing and validating the other.

In one set of experiments, Horikawa's team transplanted cells that kept producing a signal for the Notch receptor, a critical protein in many types of pattern formation. They found that the resulting embryos made smaller segments. The mathematical simulation predicted that an increase in a signal from one cell would drive its neighbours to tick faster, thereby reducing segment size. The experimental team confirmed that the phase of oscillation around the transplanted cells had indeed become quicker.

Once the researchers had validated the mathematical model experimentally, “we were able to carry out numerous virtual experiments,” says Horikawa. One of the more successful examples investigated ‘noise’ generated by erratic cell division. Using high-resolution imaging techniques, Horikawa and colleagues had found that cells actively proliferate in the synchronized oscillation zone, even though it had long been assumed that cells there do not divide. The team showed that the process of cell division changes the timing of clocks in individual cells. The mathematical model predicted that Notch signalling is critical for overcoming the effects of such noise to ensure coherent oscillations among cells... and a series of experiments confirmed this.

The imaging techniques and transplantation experiment were technically difficult and took a long time to finesse. “But the most challenging and fruitful aspect of the work was the communication between the experimental team and the theoretical team, which are very different in their cultures,” says Horikawa. “I believe that mixing these philosophies was the key to our success.”

Now such a cooperative system is in place, Horikawa hopes to delve further into the mechanism of the clock. For example, he wants to find out how the synchronized oscillation is converted into segment boundaries. “Many questions remain to be answered.”