Motivation for the model: intercellular coupling

There seems to be some evidence that PSM clock-gene oscillations are not totally cell-autonomous:

• Primitive-streak precursors of the lateral and medial parts of a same somite do not have the same anterio-posterior position in the primitive streak (Psychoyos and Stern, 1996, Selleck and Stern, 1991, Eloy-Trinquet and Nicolas, 2002), and the expression of their somitogenesis-clock genes does not oscillate in synchrony (Jouve et al., 2002). However, synchrony seems to be achieved quickly upon ingression into the PSM. While other complicated mechanisms could be at play, synchronisation of physically close cells seems to be the simplest explanation.
• Somites are not "developmental compartments": groups of PSM cells with a restricted spatial extent shortly after ingression can have descendants which span many somites (Kulesa and Fraser, 2002), and a single cell, labeled shortly before it is incorporated into a somite, can have descendants in two adjacent somites (Stern et al., 1988). It would be a complicated model, if oscillations are totally cell-autonomous, for daughter cells to inherit different phases after cell division, and to "sort out" at segmentation time.
• Inversion of the anteroposterior polarity of presumptive somites in the caudal PSM does not result in segmentation or polarity defects (Dubrulle et al., 2001); segmentation models relying on the somitogenesis clock require re-synchronisation of the inverted tissue with surrounding PSM for this to be possible.
• Analysis of zebrafish mutants in the segmentation-clock genes has led Jiang et al. (2000) to conclude that the role of the role of Notch signalling is to synchronise the segmentation clock between neighbouring cells.
• One would expect stochastic effects to have a measurable impact on the individual cellular oscillators, which could not be overcome if oscillations were cell-autonomous. Randomness in transcriptional regulation was discussed by Kepler and Elston (2001); phenotypic effects of varying biochemical parameters were shown by Ozbudak et al. (2002). Inter-cellular synchronisation was argued by Cooke (1998) to be necessary for the somitogenesis clock pattern to be so refined.