In the mass production systems of microalgal species, it is important to ensure the safety and quality of the biomass and product. This requires effective monitoring tools that are sensitive, rapid and simple to use. In this study, hyperspectral transmittance spectroscopy (HTS) was applied for the detection, cell density quantification and classification of algal and cyanobacterial species. A database of HTS data was assembled from samples of seven algal and cyanobacterial species at different cell densities and used for quantifying and classifying the species, using chemometric and machine learning algorithms. The results obtained showed the ability to quantify the species with a detection limit of 104 cells/mL for the support vector machine models applied, and classify the species at concentrations >105 cells/mL. The current study suggests that HTS is applicable for cell density quantification. HTS was used to distinguish between cell cultures of cyanobacteria and algae and was further able to distinguish between cyanobacteria species as well as algal species. In addition, reducing the dimensions (number of spectral bands) of HTS data using feature selection and aggregation improved the classification accuracy. Thus, HTS is recommended as an effective tool for monitoring and management of microalgal bioreactors.
Bibliographical noteFunding Information:
Funding for this work was provided by the Israel Ministry of Agriculture and Rural Development , Grant No # 20-06-0044 and this study was partially supported by the Hebrew University of Jerusalem 's Intramural Research Fund in Career Development and by Israel Science Foundation Grant No# 3456/20 .
© 2023 Elsevier B.V.
- Chemometric analysis
- Machine learning