Machine Learning assisted engineering of artificial microbial consortia for the delivery of bio-plastic materials from renewable waste
Researchers: Marcus Liwicki och Leonidas Matsakas
In response to the environmental challenges presented by fossil-based plastics, including energy wastage and contamination, there's a pressing need to identify biodegradable plastics derived from renewable sources. An emerging trend is the use of microbial consortia capable of metabolizing complex carbon sources, like agricultural or forest residues and industrial organic waste to produce synthetic, bio-degradable plastics. Since different source-material requires different microbial consortia, the current state of practice is to design artificial microbial consortia in costly trial-and-error (plus human experience) lab scenarios. The focus of this project is on machine learning assisted system for engineered artificial microbial consortia. They provide the potential to design microorganisms that can engage in symbiotic relationships, effectively bypassing the metabolic constraints of single strains and the inefficiencies of natural consortia. The project is expected to deliver the basis of a novel ML algorithm for assisting the engineering of artificial microbial consortia for the delivery of bio-plastic materials from renewable waste and residual feedstock.
1. Challenges
Annually, 400 million tons of plastic are produced worldwide due to its durability, broad applicability, and cost-effectiveness [1]. Predictions of the WEF indicate that by 2050, plastic consumption could quadruple[1]. This rise is set to escalate plastics’ share of global oil consumption from 6% in 2014 to an anticipated 20% by 2050, corresponding to an emission of 6.5 Gt CO2 equivalents annually [1]. Tragically, of the plastic created until 2017 (8,300 million tons) around 60% ended up in landfills or was irresponsibly discarded in the environment.
In response to the environmental challenges presented by fossil-based plastics, including energy wastage and contamination, there's a pressing need to identify biodegradable plastics derived from renewable sources [4]. Polyhydroxyalkanoates (PHAs), microorganism-produced polyesters, emerge as promising contenders, as they echo many of the favorable properties of petroleum-based plastics while being biodegradable and biocompatible [4]. However, the commercial production of PHAs is hamstrung by their production costs, mainly connected to the raw material cost [5]. Reliance on single microbial cultures narrows the spectrum of usable feedstocks (primarily to acetate and butyrate [7]), which coupled with low PHA accumulation yields [4] stands as a significant impediment to their widespread commercial adoption. The potential solution lies in microbial consortia capable of metabolizing complex carbon sources, like agricultural or forest residues and industrial organic waste. These sources are beyond the metabolic capacity of individual species [5]. While naturally occurring microbial consortia, like those found in sludge, present a viable option, they're notoriously tricky to regulate for optimal product output, leading to suboptimal process efficiency.
The focus of this project is on engineered artificial microbial consortia. They provide the potential to design microorganisms that can engage in symbiotic relationships, effectively bypassing the metabolic constraints of single strains and the inefficiencies of natural consortia. For example, engineering consortia via external stimuli has proven beneficial to the produce bio-hydrogen and platform chemicals from organic waste [2]. Yet, traditional techniques to build these artificial consortia are labor-intensive, resource-heavy, and often neglect the intricate microbial interactions, which can lead to unexpected outcomes or even consortium collapse. Thus, the capability to foresee the behavior of an artificial consortium before its introduction becomes pivotal [8].
Recent advances in high-throughput screening, sequencing, computational techniques, and machine learning (ML) have paved the way for designing, building, and analyzing complex microbial consortia [9]. This momentum has fostered a new realm in synthetic biology and biotechnology. The goal is to create artificial communities that mirror the robustness, efficiency, and flexibility of natural systems. This evolution has been propelled by high-throughput biological data sources, encompassing transcriptomics, proteomics, and metabolomics, to name a few [10]. One notable application of these techniques is the development of a machine learning-based Automated Recommendation Tool (ART) for synthetic biology. This tool offers a set of suggested strains for subsequent engineering cycles, coupled with probabilistic predictions of their production metrics [11].
From a machine learning and data science perspective, ART is very well conceptualized and easy to use, but does not include the most recent accomplishments in meta learning [12,13] and learning with sparse data (SSL – Self supervised learning) [14,15], which would be useful for scenarios where real experimental data is costly – and many particular solutions exist which have a similar high-level structure.
References
[1] Tiso, Till, et al. "The metabolic potential of plastics as biotechnological carbon sources–review and targets for the future." Metabolic engineering 71 (2022): 77-98.
[2] Sarkar O, Matsakas L, Rova U, Christakopoulos P: "Ultrasound-controlled acidogenic valorization of wastewater for biohydrogen and volatile fatty acids production: Microbial community profiling." iScience 26 (2023): 106519.
[3] Mohanan, Nisha, et al. "Microbial and enzymatic degradation of synthetic plastics." Frontiers in Microbiology 11 (2020): 580709.
[4] Zhu, Yinzhuang, Mingmei Ai, and Xiaoqiang Jia. "Optimization of a two-species microbial consortium for improved mcl-PHA production from glucose–xylose mixtures." Frontiers in Bioengineering and Biotechnology 9 (2022): 794331.
[5] Corrado, Iolanda, et al. "The power of two: An artificial microbial consortium for the conversion of inulin into Polyhydroxyalkanoates." International Journal of Biological Macromolecules 189 (2021): 494-502.
[6] Guerra-Blanco, P., et al. "Polyhydroxyalkanoates (PHA) production by photoheterotrophic microbial consortia: effect of culture conditions over microbial population and biopolymer yield and composition." European Polymer Journal 98 (2018): 94-104.
[7] Morgan-Sagastume, F., et al. "Polyhydroxyalkanoate (PHA) production from sludge and municipal wastewater treatment." Water Science and Technology 69.1 (2014): 177-184.
[8] Karkaria, Behzad D., Alex JH Fedorec, and Chris P. Barnes. "Automated design of synthetic microbial communities." Nature communications 12.1 (2021): 672.
[9] Rapp, Kent M., Jackson P. Jenkins, and Michael J. Betenbaugh. "Partners for life: building microbial consortia for the future." Current Opinion in Biotechnology 66 (2020): 292-300.
[10] Dasgupta et al., Comput Methods Programs Biomed 2020;192:105436
[11] Radivojević, Tijana, et al. "A machine learning Automated Recommendation Tool for synthetic biology." Nature communications 11.1 (2020): 4879.
[12] Huisman, Mike, Jan N. Van Rijn, and Aske Plaat. "A survey of deep meta-learning." Artificial Intelligence Review 54.6 (2021): 4483-4541.
[13] Upadhyay, R., Phlypo, R., Saini, R., & Liwicki, M. (2021). Sharing to learn and learning to share--Fitting together Meta-Learning, Multi-Task Learning, and Transfer Learning: A meta review. arXiv preprint arXiv:2111.12146.
[14] Jaiswal, Ashish, et al. "A survey on contrastive self-supervised learning." Technologies 9.1 (2020): 2.
[15] Chhipa, P. C., Upadhyay, R., Pihlgren, G. G., Saini, R., Uchida, S., & Liwicki, M. "Magnification prior: a self-supervised method for learning representations on breast cancer histopathological images." Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. 2023.
Updated: