Machine learning of ramanomes helps screening microalgal cell factories that fix carbon dioxide. Credit: LIU Yang
Researchers from the Qingdao Institute of Bioenergy and Bioprocess Technology (QIBEBT) of the Chinese Academy of Sciences (CAS) have developed a way to rapidly determine exactly which microalgae can most readily convert carbon dioxide into valuable compounds that can be used for fuels, food and drugs. They published their approach on June 18 in Analytical Chemistry.
The current approach to identifying microalgae and understanding their metabolic activity involves culturing and studying each species. “It is slow and tedious,” said first author Mohammadhadi Heidari Baladehi, a doctoral student in the Single-Cell Center and CAS Key Laboratory of Biofuels at QIBEBT. “Moreover, the vast majority of microalgae in nature are not yet cultured.”
To speed up the assessment, researchers employed Raman Microspectroscopy, which produces images that reveal the metabolic activities of the cell.
A machine learning approach for the ramanome database
In this work, Heidari Baladehi and his team established a database of “ramanomes,” a collection of single-cell Raman spectra of more than 9,000 known, diverse microalgae species. To demonstrate the power of the database they applied a machine learning approach, meaning, as more information was added to the system, the more the system learned to identify functional and genetic patterns between different organisms.
Heidari Baladehi said a critical strength of their approach was combining two Raman-based “portraits,” one for pigments and one for all the other compounds in the cell. Most current approaches collect only one portrait, and usually do not collect them from the same cell. The QIBEBT researchers proposed to combine the two portraits, so that a much more complete and richer range of information can be gleaned. Given the combined portraits and the machine learning algorithm, their system can identify species and their metabolic functions with 97% accuracy, for those microalgae already cultured and recorded.
For those species that have not been cultured, the QIBEBT researchers invented a different strategy: the cells are imaged for the Raman portraits to profile their metabolic functions first, and then sorted and sequenced for the genome sequences, one cell at a time. They achieved this using an instrument developed at Single-Cell Center called RACS-Seq. The instrument is unique in its ability to produce high-quality genome sequences for the target cell, at the resolution of precisely one cell, after collecting the Raman signal.
“This comprehensive approach for rapidly identifying and metabolically profiling single-cells, either cultured or uncultured, greatly accelerates the mining and screening of microalgal cell factories for carbon-neutral production,” said XU Jian, Director of Single-Cell Center and a senior author of the study.
Based on their Microalgal Ramanome Database, the researchers have established an open-access web platform to support the function-based ID system. They plan to further develop their system to accommodate other classes of living organisms on Earth.
All rights reserved. Permission required to reprint articles in their entirety. Must include copyright statement and live hyperlinks. Contact david@algaeplanet.com. Algae Planet accepts unsolicited manuscripts for consideration, and takes no responsibility for the validity of claims made in submitted editorial.
