Phase II Commercialization of HABStats: A Harmful Algal Bloom Species, Toxin and Total Phytoplankton Identification System

In Phase I, we developed a prototype instrument for counting, sizing and identifying both non-toxic and toxic Harmful Algal Bloom (HAB) species. We did this by integrating three innovative technologies: High-throughput color, Light Field Imaging Flow Cytometry to acquire morphological information for each cell, Raman spectroscopy to acquire species-specific molecular information on pigments and triglycerides, and Convolutional Deep Neural Networks providing Deep Learning data integration and an artificial taxonomist. We showed that at least 10 to 100 cells per second could be processed in real-time. We built an image and Raman library of more than 22 marine and 30 freshwater HAB species. Several HAB toxins were characterized and quantified by Raman spectroscopy. Phase II Objectives: continue building Raman libraries focusing on HAB toxins; examine the specificity of Raman compared to HPLC and ELISA analysis of HAB toxins; complete integration of 3-D imaging, Raman spectroscopy, and Deep Learning classification of cells and toxins into a Gen2 prototype that may be provided to beta-testers; work with Lockheed Martin to develop a Design for Manufacturing Plan for the HABStats instrument; construct four Gen2 HABStats and provide them to beta-testers involved in management and research of both marine and freshwater HABs.