Parameter Identification in Reaction-Diffusion-Chemotaxis Systems from Single Turing Pattern Amplitudes

Abstract

Turing patterns encode key information about biological mechanisms, yet traditional inverse problems rely on non-biological data like boundary measurements, neglecting the patterns themselves. We introduce a new direction that directly uses the amplitudes of Turing patterns for parameter identification. Motivated by stripe-like biological patterns, we study two 1D models: one with density-dependent chemotaxis ($\chi(n,c)=\chi_0 n$) and one with ratio-dependent chemotaxis ($\chi(n,c)=\chi_0 n/c$). We present a framework that uses the spatial amplitude profile of a stationary pattern to recover system parameters, including wavelength, diffusion constants, and chemotactic and kinetic coefficients---offering a biologically grounded paradigm for reverse-engineering pattern formation. The core contribution is a proof-of-concept showing that, under a Fourier truncation at mode $M=3$, the amplitude data can reconstruct the parameter combinations $\{d_n k^2,\, d_c k^2,\, \chi_0 k^2,\, r\}$ for both models. These uniquely determine the ratios $d_c/d_n$ and $\chi_0/d_n$, while individual parameters are recovered up to an overall scale. This is a purely theoretical work; numerical validation and stability analysis are left for future research. The framework is extensible to other models with suitable modifications.