Automated design of synthetic microbial communities

Abstract

<jats:title>Abstract</jats:title><jats:p>In naturally occurring microbial systems, species rarely exist in isolation. There is strong ecological evidence for a positive relationship between species diversity and the functional output of communities. The pervasiveness of these communities in nature highlights that there may be advantages for engineered strains to exist in cocultures as well. Building synthetic microbial communities allows us to create distributed systems that mitigates issues often found in engineering a monoculture, especially when functional complexity is increasing. Here, we demonstrate a methodology for designing robust synthetic communities that use quorum sensing to control amensal bacteriocin interactions in a chemostat environment. We explore model spaces for two and three strain systems, using Bayesian methods to perform model selection, and identify the most robust candidates for producing stable steady state communities. Our findings highlight important interaction motifs that provide stability, and identify requirements for selecting genetic parts and tuning the community composition.</jats:p>

Single strain control of microbial consortia

“…We can assess the candidate models’ ability to produce stable consortia using approximate Bayesian computation sequential Monte Carlo (ABC SMC) (Fig. 9 b), which allows us to approximate model and parameter posterior probabilities by random sampling and weight assignment through a series of intermediate distributions 45 – 47 . The output of ABC SMC is an approximation of the posterior distribution of models and of the parameters for each model.…”

“…9 remove the requirement for competitive exclusion by the addition of self-limitation. We have previously discovered such motifs to improve the robustness of community control architectures 45 .…”

“…Future work will involve the implementation of the system in more complex communities 48 , 49 and in animal models including C. elegans 50 . In previous work, we have shown mathematically that bacteriocins can be used to control more complex consortia 45 . Extensions of the system demonstrated here with additional bacteriocins and quorum molecules could allow its use in industrial biotechnology for the control of communities in bioreactors in which division-of-labour is desirable to prevent burden.…”

“…The model exploration using ABC SMC 45 is available at https://github.com/ucl-cssb/AutoCD . Plate reader and flow cytometry data was processed using FlopR v0.3.0 40 , which is available at https://github.com/ucl-cssb/flopr .…”

MiMiC: a bioinformatic approach for generation of synthetic communities from metagenomes

“…Moreover, communities of surprisingly low complexity were inferred using random selection of metabolite pairs (substrate product), and no information was provided on the complexity of usage and computation time. The second method aimed to generate communities that would be stable within a chemostat based on a user-provided list of strains with known quorum sensing and bacteriocin production and sensitivity values (Karkaria et al, 2021). This method was designed for application to engineered strains, for which such data would be known due to having been designed into the strains genome, and not for environmental isolates.…”