Preprint: a neural network inferred how plant clocks respond to darkness from standard recordings
Preprint: a neural network inferred how plant clocks respond to darkness from standard recordings
A study posted on the bioRxiv preprint server on October 24 describes a method that uses observations of plants exposed to two hours of darkness to estimate how their internal clocks change speed at different times of day. The authors tested the method on recordings from Arabidopsis thaliana, a common laboratory plant.
Circadian clocks generate an internal cellular rhythm with a period of approximately one day. The same two-hour exposure to darkness shifts this rhythm differently depending on when in the cycle it begins. A conventional phase response curve shows the resulting phase shift. The authors use the term infinitesimal phase response curve, or iPRC, for a function that describes how a stimulus changes the speed of the clock during exposure. The model uses this function to calculate the response to a different schedule of darkness.
The authors inferred this function from a set of standard recordings. They observed plants given the same two-hour pulse of darkness at different points in their internal cycle and tracked their rhythms through light emitted by an inserted reporter gene. One part of the model describes how phase changes over time, while another fits a sensitivity function that repeats each day. A phase equation links the two: the sensitivity function, combined with the known schedule of darkness, must explain the observed change in phase.
In a 2021 study, researchers had already reconstructed the phase response curve of the Arabidopsis thaliana clock by deliberately desynchronizing the plant's cellular clocks and then applying a single pulse. The new method uses recordings from plants with synchronized rhythms.
The method was first tested on computer models of oscillations for which the correct sensitivity function was known in advance. It was then applied to recordings from three laboratory lines of Arabidopsis thaliana with altered clock genes: prr5, prr7 and prr9. The shapes of the functions and the times of greatest sensitivity were similar to those obtained in the dedicated 2021 experiment.
The authors used the inferred functions to calculate responses to 8, 12 and 16 hours of darkness. For prr7, the model predicted that the peak would occur at nearly the same time after darkness ended. In independent recordings from roots of another line, prr7-11, peak timing also depended only weakly on the duration of darkness. This was the feature the authors compared.
The method uses recordings of a two-hour pulse to predict the response to a longer period of darkness, which can be tested in a separate measurement.