Bioengineering allows experts to design biological systems that benefit society, such as cells that produce biofuels or drugs that fight antibiotic-resistant infections. This is achieved by analyzing and redesigning the series of chemical reactions that produce a specific compound in a cell.
However, reliable prediction of the outcome of Bioengineering work is difficult because the traditional mathematical models used to foresee the dynamics of chemical reactions in a cell take months to develop and require a high level of expertise in this field. Moreover, these predictions do not always match experimental results.
The external scientific member of the Basque Center for Applied Mathematics – BCAM, and regular collaborator of the center, Héctor García has recently published the article “A machine learning approach to predict metabolic pathway dynamics from time-series multiomics data” in the journal Nature Systems Biology and Applications suggesting that a new approach to Machine Learning could speed up the processes of Bioengineering.
Machine Learning uses data to train a computer algorithm to make predictions. The algorithm learns the behavior of a system by analyzing data from related systems, allowing scientists to quickly predict the results of a process.
Héctor García and Zak Costello, his partner in this research project carried out at Lawrence Berkeley National Laboratory (Berkeley Lab), have developed artificial intelligence techniques that are able to predict and design Bioengineering processes more quickly. They claim that these Machine Learning techniques can replace traditional kinetic models because they are more effective, since the system is able to learn directly from data and examples and does not require a biological expert.
As Garcia states in a press released published in Berkeley Lab, their approach promises to accelerate the development of biomolecules for many applications in addition to commercially viable biofuels, such as antibiotic-resistant infection-fighting drugs and drought-resistant crops, and could become “a new way to guide Bioengineering efforts”.
García will visit BCAM next July to collaborate with the Life and Materials Science Modelling and Simulation group led by the Ikerbasque Research Professor Elena Akhmatskaya.