We developed an #informationtheory method, MITE, to identify functional cell-cell connections by observing the coevolution of circadian gene expression as the cells synchronized. This framework can be used for inferring connectivity in any cellular network. (3/n)
Mammalian clock is efficiently wired: SCN connectivity patterns were highly conserved, both bilaterally and across mice. SCN is sparsely connected compared to other brain areas (~720 connections/cell), yet signals reached any other SCN cell via just 3–4 intermediaries! (4/n)
While SCN cells connect as a #smallworld network with most cells signaling proximally; intriguingly, we found a few long-range connections from ventral cells that project halfway > across the SCN, short-circuiting the network to enhance signal transmission. (5/n)
Two asymmetrically coupled cellular networks drive SCN synchrony: Within each explant, we found dorsal and ventral cellular modules resembling the anatomical shell-core regions. Notably, these were absent in explants that failed to synchronize during our experiments. (6/n)
The SCN exhibited strong ventral-to-dorsal signaling, with over 60% of all connections originating from ventral cells. Accordingly, ventral module cells synchronize their rhythms first, driving the rest of the network. (7/n)
We identified four functional cell types that mediate circadian signaling across the SCN neural network, acting as circadian signal 'generators,' 'broadcasters,' 'bridges,' or 'sinks.' (8/n)
I've had this discussion with Lance Riley :) It is a very interesting system! If you can extract time series/pixel intensities from each nucleus, inter-nucleus communication can be mapped with this method.
I've developed a python program for circadian analysis at pixel resolution in tissue recordings. Happy to share it. We have used it in this work - https://www.pnas.org/doi/10.1073/pnas.2400339121
Comments
https://www.pnas.org/doi/10.1073/pnas.2400339121