“Applications of AI in Plant cell metabolic engineering”

   Present climate change scenario in a rapidly growing population threatens the food and nutritional security of the world. To achieve food security tremendous work has been accomplished in traditional breeding to increase the crop yield. However, loss in biodiversity has resulted in an imbalance of nutrients, risking deficiency of nutrients in consumers. FAO has reported that approximately two billion people around the world suffer from “hidden hunger”. Application of traditional breeding for improving nutritional quality is limited due to long breeding cycles and limited nutrient variability. This necessitates a paradigm shift in the crop improvement approach towards metabolic engineering, which involves identifying key pathways and enzymes that can be modified for the optimal production of the target molecule.

To utilize metabolic engineering to its maximum extent a high throughput method or tool that characterizes cells, develops cell-specific metabolic maps and link the maps of cell populations so that engineering strategies can be devised. AI (Artificial Intelligence) is a tool that has been utilized in various fields to improve its efficiency.

Through the integration of AI, will be able to address the limitations in metabolic engineering through its application in pathway design where RetroPath RL, a retrosynthetic method that implemented reinforcement learning, using the reaction rules devised for enzymatic reactions with Monte-Carlo tree search algorithm to plan the biosynthetic pathway for a target chemical1. Pathway optimization or metabolic flux optimization can be achieved by identifying the optimal combination of multiple gene expression levels within a pathway to maximize product titers, rates and yields (TRY) which are accomplished by implementing active learning and Bayesian optimization2,3. Gene annotation utilizing AI helps in the large-scale identification of genes, their sequences, length and structure and therefore, helps in finding alternative organisms where the same gene, pathways or gene clusters exist4.

Artificial intelligence provides an opportunity to make metabolic engineering more predictable and efficient. The implementation of AI in plant cell metabolic engineering has the potential to revolutionize the way to approach crop improvement and address the nutritional challenges of the modern world.

References: 

1. KOCH M, DUIGOU T AND FAULON J. L., 2019, Reinforcement learning for bioretrosynthesis. ACS Synth. Biol., 9 (1): 157–168. 

2. OPGENORTH P, COSTELLO Z, OKADA T, GOYAL, G, CHEN, Y, GIN, J AND BELLER H R, 2019, Lessons from two design–build–test–learn cycles of dodecanol production in Escherichia coli aided by machine learning, ACS Synth. Biol., 8: 1337–1351. 

3. HAMEDIRAD M, CHAO , WEISBERG S, LIAN J, SINHA S AND ZHAO H, 2019, Towards a fully automated algorithm driven platform for biosystems design. Nat. commun., 10(1): 5150-5160. 4. AMIN M R, YUROVSKY A, TIAN Y AND SKIENA S, 2018, Deep annotator: Genome Annotation with Deep Learning. In Proceedings of the 2018 ACM international conference on bioinformatics, computational biology and health informatics pp. 254-259.