日時：	15:00 - 16:30, Aug 7 (Wed), 2024
場所：	Room 501, 5th floor, Bldg. #3 Institute for Life and Medical Sciences, Kyoto University
演者：	Dr. Kento Nakamura　(RIKEN)
演題：	Optimal control of run‐and‐tumble chemotaxis under sensory noise and its biochemical implementation

講演要旨

Abstract: Various organisms use run-and-tumble chemotaxis to search for food and
desired environments. In this search strategy, organisms climb the spatial gradient of
ligand concentration by controlling the frequency of random directional changes
(tumble rate) based on temporal changes in ligand concentration sensed during
ballistic swimming (run). Previous studies have investigated how efficient the runand-
tumble chemotaxis can be in principle by deriving optimal control of tumble rate
and comparing it with actual motor responses. However, these derivations ignore
sensory noise and nonlinearity of response due to theoretical difficulties, hindering
the detailed correspondence between optimal control and biochemical systems.

In this talk, we address this gap by deriving the nonlinear optimal control of run-andtumble
motion under sensory noise based on partially observable stochastic control
and Kullback-Leibler control theory [1]. The derived optimal control consists of
nonlinear filtering dynamics that estimates swimming direction from noisy signals and
an optimal control function that maps this estimate to tumble rate. Notably, this
optimal control reproduces a standard biochemical model of Escherichia coli’s
sensory system and experimentally observed nonlinear response relations. This
finding suggests that E. coli’s biochemical pathway may implement an efficient
structure for run-and-tumble chemotaxis.

Reference:
[1] Nakamura, K., & Kobayashi, T. J. (2022). Optimal sensing and control of run-andtumble
chemotaxis. Physical Review Research, 4,
013120. [https://doi.org/10.1103/PhysRevResearch.4.013120]

 

contact: Takashi Okada
Laboratory of Mathematical Biology,
Institute for Life and Medical Sciences, Kyoto University
Email: okada.Takashi.3z＊infront.kyoto-u.ac.jp（＊→＠へ変えてください）