Engineering Deep, Synthetic Logic Networks in Bacteria for Generating Temporal Encodings of Dynamic Biochemical Cues
A) A set of potential time-course profiles (a step, square wave, and a Gaussian) of DoD relevant biochemicals that will be (B) used to induce the deep, synthetic logic bifan network which transforms the signal through each layer in the network with Hill function (sigmoid-like) activation functions. (C) The information processing outcome of each layer can be viewed as a temporal encoding wherein each layer scales then nonlinearly transforms the incoming signals such that the output profiles are a complex representation of the inputs in a potentially new topological space. These temporal encodings can then be deconvolved to obtain the raw input signal that produced the temporal encodings, with aid from machine learning algorithms trained on input-output response profiles.
This proposal aims to develop an integrated experimental-computational framework for engineering deep, synthetic logic networks in bacteria for encoding and extracting temporal data for DoD relevant biochemical signals.
We seek to construct multi-layer deep logic networks for encoding detailed temporal sequence information of a biochemical analyte. We will construct cellular logic networks with up to seven layers, enacting XOR logic to encode and pass signals from one layer to the next, like an artificial neural network. Each layer will act as a temporal filter or transfer function, to encode delay, while simultaneously projecting a temporal signal onto a nonlinear function space for each cell. The composition of multiple layers will create, in aggregate, a nonlinear encoding function that transforms a temporal sequence of time-varying biochemical signal or concentration into output kinetics for each layer in the network, which can be read directly by fluorometric measurements or a few static RT-qPCR measurements. In this way, we will develop a new class of integrated biotechnology with inverse modeling tools for encoding DoD relevant temporal signals in the environment.
