Closing the loop on morphogenesis: a mathematical model of morphogenesis by closed-loop reaction-diffusion
As noted in the introduction, measuring waves is technically challenging. While waves have been observed the lab, confirming our hypothesis would be much easier if future work creates transgenics with fluorescent reporters that allow readouts of the morphological computation in the living state.
We have demonstrated a closed-loop negative-feedback machine to control morphogenesis, as a contribution to the efforts to understand morphogenesis as a target-directed process (Levin et al., 2023). Its goal is to lay down N copies (for a reasonably arbitrary N) of a simple RD pattern; e.g., to be used as repeats of a coordinate system. It achieves this goal with a closed-loop negative-feedback controller that• employs waves to count morphogen peaks, and thus count the current number of pattern repetitions. The waves work by taking advantage of existing asymmetry.• compares the current number of peaks to its goal N. • adjusts the RD pattern-length parameter λRD so as to move the pattern towards the goal.
By combining these concepts, we have• created multiple-peak RD patterns more reliably than previous work• as such, laid down multiple gradients for PI acting downstream of RD• controllably made a subset of the pattern segments larger or smaller than the others, so as to subdivide a field into unequal subfields in a repeatable manner.
The circuit described here enables flexible actions under a range of circumstances to reach a specific large-scale goal state which belongs to the collective and not the individuals (Levin, 2019; Levin, 2022). Such capacity has been proposed as a definition of intelligence (Fields and Levin, 2022), for the field of basal cognition (Lyon, 2006; Lyon, 2015; Sole et al., 2016; Urrios et al., 2016; Macia et al., 2017). Thus, the above circuit and analysis not only supports a way of viewing morphogenetic processes as a set of specific computational tasks, but also suggests a synthetic biology architecture for incorporating a simple kind of intelligence into novel biological constructs (Doursat et al., 2013; Doursat and Sanchez, 2014; Kamm and Bashir, 2014; Kamm et al., 2018; Davies and Glykofrydis, 2020; Glykofrydis et al., 2021; Davies and Levin, 2022). Future work will investigate the presence of these dynamics in vivo, as well as use the insights revealed by this modeling process to create novel patterning systems for synthetic biorobotics, regenerative medicine, and tissue bioengineering.