For the first time, a Newcastle University study has demonstrated that machine learning can predict the biological characteristics of Rubisco, the most prevalent enzyme on Earth.
Carbon for practically all life on Earth is provided via carboxylase/oxygenase. In order to maintain the majority of life on Earth, Rubisco converts atmospheric CO2 from the planet’s atmosphere into organic carbon matter.
Since there is natural variety among the Rubisco proteins of land plants, agricultural plants can absorb and retain more atmospheric CO2 when Rubisco proteins with certain functional qualities are transplanted, according to modelling studies.
Wasim Iqbal, the study’s primary author and a Ph.D. researcher at Newcastle University’s School of Natural and Environmental Sciences in Dr. Maxim Kapralov’s lab, created a machine learning tool that has a surprising high degree of accuracy in predicting the performance characteristics of many land plant Rubisco proteins. With the help of this technology, researchers hope to find a “supercharged” Rubisco protein that can be bioengineered into important crops like wheat.
A valuable method for screening and forecasting plant Rubisco kinetics for engineering projects as well as for fundamental research on Rubisco evolution and adaptation is presented in the article, which was published in the Journal Of Experimental Botany. The main method utilised to locate improved Rubiscos for crop engineering efforts is screening the natural diversity of Rubisco kinetics.
Their research, according to Wasim, “will have enormous consequences for climate models and bioengineered crops.”
This discovery gives plant scientists a pre-screening method to identify Rubisco species with better kinetics for crop engineering initiatives.
The use of machine learning techniques can improve the accuracy of global photosynthesis estimations. Earth system models (ESM) utilised by climate scientists are compatible with the Rubisco performance characteristics predicted by our model. Currently, ESMs estimate photosynthesis at the ecosystem scale using a single set of Rubisco parameters from the same species (or occasionally a few). Our machine learning technique could provide predictions for the majority of terrestrial plants, enhancing the ESMs’ accuracy.”
The following steps in this research involve extracting the top Rubisco proteins found through laboratory predictions and attempting to bioengineer a plant species using a foreign Rubisco protein.
