Linear Noise Approximation of Intensity-Driven Signal Transduction Channels (1908.10897v1)

Abstract: Biochemical signal transduction, a form of molecular communication, can be modeled using graphical Markov channels with input-modulated transition rates. Such channel models are strongly non-Gaussian. In this paper we use a linear noise approximation to construct a novel class of Gaussian additive white noise channels that capture essential features of fully- and partially-observed intensity-driven signal transduction. When channel state transitions that are sensitive to the input signal are directly observable, high-frequency information is transduced more efficiently than low-frequency information, and the mutual information rate per bandwidth (spectral efficiency) is significantly greater than when sensitive transitions and observable transitions are disjoint. When both observable and hidden transitions are input-sensitive, we observe a superadditive increase in spectral efficiency.