Distributed classifier based on genetically engineered bacterial cell cultures

Biological computation through recurrence

One of the defining features of living systems is their adaptability to changing environmental conditions. This requires organisms to extract temporal and spatial features of their environment, and use that information to compute the appropriate response. In the last two decades, a growing body of work, mainly coming from the machine learning and computational neuroscience fields, has shown that such complex information processing can be performed by recurrent networks. Temporal computations arise in these networks through the interplay between the external stimuli and the network's internal state. In this article we review our current understanding of how recurrent networks can be used by biological systems, from cells to brains, for complex information processing. Rather than focusing on sophisticated, artificial recurrent architectures such as long short-term memory (LSTM) networks, here we concentrate on simpler network structures and learning algorithms that can be expected to have been found by evolution. We also review studies showing evidence of naturally occurring recurrent networks in living organisms. Lastly, we discuss some relevant evolutionary aspects concerning the emergence of this natural computation paradigm.

Connectome-based reservoir computing with the conn2res toolbox

The connection patterns of neural circuits form a complex network. How signaling in these circuits manifests as complex cognition and adaptive behaviour remains the central question in neuroscience. Concomitant advances in connectomics and artificial intelligence open fundamentally new opportunities to understand how connection patterns shape computational capacity in biological brain networks. Reservoir computing is a versatile paradigm that uses high-dimensional, nonlinear dynamical systems to perform computations and approximate cognitive functions. Here we present conn2res: an open-source Python toolbox for implementing biological neural networks as artificial neural networks. conn2res is modular, allowing arbitrary network architecture and dynamics to be imposed. The toolbox allows researchers to input connectomes reconstructed using multiple techniques, from tract tracing to noninvasive diffusion imaging, and to impose multiple dynamical systems, from spiking neurons to memristive dynamics. The versatility of the conn2res toolbox allows us to ask new questions at the confluence of neuroscience and artificial intelligence. By reconceptualizing function as computation, conn2res sets the stage for a more mechanistic understanding of structure-function relationships in brain networks.

Synthetic nonlinear computation for genetic circuit design

Computation frameworks have been studied in synthetic biology to achieve biosignals integration and processing, for biosensing and therapeutics applications. Biological systems exhibit nonlinearity across scales from the molecular level, to biochemical network and intercellular systems. At the molecular level, cooperative bindings contribute to nonlinear molecular signal processing in a way similar to weight variables. At the intracellular network level, feedback and feedforward regulations result in cell behaviors such as multistability and adaptation. When biochemical networks are distributed in different cell groups, intercelluar networks can generate population dynamics. Here, we review works that highlight nonlinear computations in synthetic biology. We group the works according to the scale of implementations, from the cis-transcription level, to biochemical circuit level and cellular networks.

Synthetic neural-like computing in microbial consortia for pattern recognition

Complex biological systems in nature comprise cells that act collectively to solve sophisticated tasks. Synthetic biological systems, in contrast, are designed for specific tasks, following computational principles including logic gates and analog design. Yet such approaches cannot be easily adapted for multiple tasks in biological contexts. Alternatively, artificial neural networks, comprised of flexible interactions for computation, support adaptive designs and are adopted for diverse applications. Here, motivated by the structural similarity between artificial neural networks and cellular networks, we implement neural-like computing in bacteria consortia for recognizing patterns. Specifically, receiver bacteria collectively interact with sender bacteria for decision-making through quorum sensing. Input patterns formed by chemical inducers activate senders to produce signaling molecules at varying levels. These levels, which act as weights, are programmed by tuning the sender promoter strength Furthermore, a gradient descent based algorithm that enables weights optimization was developed. Weights were experimentally examined for recognizing 3 × 3-bit pattern.

Reinforcement learning in synthetic gene circuits

Synthetic gene circuits allow programming in DNA the expression of a phenotype at a given environmental condition. The recent integration of memory systems with gene circuits opens the door to their adaptation to new conditions and their re-programming. This lays the foundation to emulate neuromorphic behaviour and solve complex problems similarly to artificial neural networks. Cellular products such as DNA or proteins can be used to store memory in both digital and analog formats, allowing cells to be turned into living computing devices able to record information regarding their previous states. In particular, synthetic gene circuits with memory can be engineered into living systems to allow their adaptation through reinforcement learning. The development of gene circuits able to adapt through reinforcement learning moves Sciences towards the ambitious goal: the bottom-up creation of a fully fledged living artificial intelligence.

From Microbial Communities to Distributed Computing Systems

A distributed biological system can be defined as a system whose components are located in different subpopulations, which communicate and coordinate their actions through interpopulation messages and interactions. We see that distributed systems are pervasive in nature, performing computation across all scales, from microbial communities to a flock of birds. We often observe that information processing within communities exhibits a complexity far greater than any single organism. Synthetic biology is an area of research which aims to design and build synthetic biological machines from biological parts to perform a defined function, in a manner similar to the engineering disciplines. However, the field has reached a bottleneck in the complexity of the genetic networks that we can implement using monocultures, facing constraints from metabolic burden and genetic interference. This makes building distributed biological systems an attractive prospect for synthetic biology that would alleviate these constraints and allow us to expand the applications of our systems into areas including complex biosensing and diagnostic tools, bioprocess control and the monitoring of industrial processes. In this review we will discuss the fundamental limitations we face when engineering functionality with a monoculture, and the key areas where distributed systems can provide an advantage. We cite evidence from natural systems that support arguments in favor of distributed systems to overcome the limitations of monocultures. Following this we conduct a comprehensive overview of the synthetic communities that have been built to date, and the components that have been used. The potential computational capabilities of communities are discussed, along with some of the applications that these will be useful for. We discuss some of the challenges with building co-cultures, including the problem of competitive exclusion and maintenance of desired community composition. Finally, we assess computational frameworks currently available to aide in the design of microbial communities and identify areas where we lack the necessary tools.

Recent advances in physical reservoir computing: A review

Reservoir computing is a computational framework suited for temporal/sequential data processing. It is derived from several recurrent neural network models, including echo state networks and liquid state machines. A reservoir computing system consists of a reservoir for mapping inputs into a high-dimensional space and a readout for pattern analysis from the high-dimensional states in the reservoir. The reservoir is fixed and only the readout is trained with a simple method such as linear regression and classification. Thus, the major advantage of reservoir computing compared to other recurrent neural networks is fast learning, resulting in low training cost. Another advantage is that the reservoir without adaptive updating is amenable to hardware implementation using a variety of physical systems, substrates, and devices. In fact, such physical reservoir computing has attracted increasing attention in diverse fields of research. The purpose of this review is to provide an overview of recent advances in physical reservoir computing by classifying them according to the type of the reservoir. We discuss the current issues and perspectives related to physical reservoir computing, in order to further expand its practical applications and develop next-generation machine learning systems.

An Orthogonal Multi-input Integration System to Control Gene Expression in Escherichia coli

In many biotechnological applications, it is useful for gene expression to be regulated by multiple signals, as this allows the programming of complex behavior. Here we implement, in Escherichia coli, a system that compares the concentration of two signal molecules, and tunes GFP expression proportionally to their relative abundance. The computation is performed via molecular titration between an orthogonal σ factor and its cognate anti-σ factor. We use mathematical modeling and experiments to show that the computation system is predictable and able to adapt GFP expression dynamically to a wide range of combinations of the two signals, and our model qualitatively captures most of these behaviors. We also demonstrate in silico the practical applicability of the system as a reference-comparator, which compares an intrinsic signal (reflecting the state of the system) with an extrinsic signal (reflecting the desired reference state) in a multicellular feedback control strategy.

Synthetic Gene Circuits Learn to Classify

An efficient computational algorithm is developed to design microRNA-based synthetic cell classifiers and to optimize their performance.

Artificial Symmetry-Breaking for Morphogenetic Engineering Bacterial Colonies

Morphogenetic engineering is an emerging field that explores the design and implementation of self-organized patterns, morphologies, and architectures in systems composed of multiple agents such as cells and swarm robots. Synthetic biology, on the other hand, aims to develop tools and formalisms that increase reproducibility, tractability, and efficiency in the engineering of biological systems. We seek to apply synthetic biology approaches to the engineering of morphologies in multicellular systems. Here, we describe the engineering of two mechanisms, symmetry-breaking and domain-specific cell regulation, as elementary functions for the prototyping of morphogenetic instructions in bacterial colonies. The former represents an artificial patterning mechanism based on plasmid segregation while the latter plays the role of artificial cell differentiation by spatial colocalization of ubiquitous and segregated components. This separation of patterning from actuation facilitates the design-build-test-improve engineering cycle. We created computational modules for CellModeller representing these basic functions and used it to guide the design process and explore the design space in silico. We applied these tools to encode spatially structured functions such as metabolic complementation, RNAPT7 gene expression, and CRISPRi/Cas9 regulation. Finally, as a proof of concept, we used CRISPRi/Cas technology to regulate cell growth by controlling methionine synthesis. These mechanisms start from single cells enabling the study of morphogenetic principles and the engineering of novel population scale structures from the bottom up.

Synthetic biology routes to bio-artificial intelligence

The design of synthetic gene networks (SGNs) has advanced to the extent that novel genetic circuits are now being tested for their ability to recapitulate archetypal learning behaviours first defined in the fields of machine and animal learning. Here, we discuss the biological implementation of a perceptron algorithm for linear classification of input data. An expansion of this biological design that encompasses cellular 'teachers' and 'students' is also examined. We also discuss implementation of Pavlovian associative learning using SGNs and present an example of such a scheme and in silico simulation of its performance. In addition to designed SGNs, we also consider the option to establish conditions in which a population of SGNs can evolve diversity in order to better contend with complex input data. Finally, we compare recent ethical concerns in the field of artificial intelligence (AI) and the future challenges raised by bio-artificial intelligence (BI).

Synthetic biology approaches in cancer immunotherapy, genetic network engineering, and genome editing

Investigations into cells and their contents have provided evolving insight into the emergence of complex biological behaviors. Capitalizing on this knowledge, synthetic biology seeks to manipulate the cellular machinery towards novel purposes, extending discoveries from basic science to new applications. While these developments have demonstrated the potential of building with biological parts, the complexity of cells can pose numerous challenges. In this review, we will highlight the broad and vital role that the synthetic biology approach has played in applying fundamental biological discoveries in receptors, genetic circuits, and genome-editing systems towards translation in the fields of immunotherapy, biosensors, disease models and gene therapy. These examples are evidence of the strength of synthetic approaches, while also illustrating considerations that must be addressed when developing systems around living cells.

Nano-guided cell networks as conveyors of molecular communication

Advances in nanotechnology have provided unprecedented physical means to sample molecular space. Living cells provide additional capability in that they identify molecules within complex environments and actuate function. We have merged cells with nanotechnology for an integrated molecular processing network. Here we show that an engineered cell consortium autonomously generates feedback to chemical cues. Moreover, abiotic components are readily assembled onto cells, enabling amplified and 'binned' responses. Specifically, engineered cell populations are triggered by a quorum sensing (QS) signal molecule, autoinducer-2, to express surface-displayed fusions consisting of a fluorescent marker and an affinity peptide. The latter provides means for attaching magnetic nanoparticles to fluorescently activated subpopulations for coalescence into colour-indexed output. The resultant nano-guided cell network assesses QS activity and conveys molecular information as a 'bio-litmus' in a manner read by simple optical means.

Multi-input distributed classifiers for synthetic genetic circuits

For practical construction of complex synthetic genetic networks able to perform elaborate functions it is important to have a pool of relatively simple modules with different functionality which can be compounded together. To complement engineering of very different existing synthetic genetic devices such as switches, oscillators or logical gates, we propose and develop here a design of synthetic multi-input classifier based on a recently introduced distributed classifier concept. A heterogeneous population of cells acts as a single classifier, whose output is obtained by summarizing the outputs of individual cells. The learning ability is achieved by pruning the population, instead of tuning parameters of an individual cell. The present paper is focused on evaluating two possible schemes of multi-input gene classifier circuits. We demonstrate their suitability for implementing a multi-input distributed classifier capable of separating data which are inseparable for single-input classifiers, and characterize performance of the classifiers by analytical and numerical results. The simpler scheme implements a linear classifier in a single cell and is targeted at separable classification problems with simple class borders. A hard learning strategy is used to train a distributed classifier by removing from the population any cell answering incorrectly to at least one training example. The other scheme implements a circuit with a bell-shaped response in a single cell to allow potentially arbitrary shape of the classification border in the input space of a distributed classifier. Inseparable classification problems are addressed using soft learning strategy, characterized by probabilistic decision to keep or discard a cell at each training iteration. We expect that our classifier design contributes to the development of robust and predictable synthetic biosensors, which have the potential to affect applications in a lot of fields, including that of medicine and industry.