Instrument and Method Development for Single-Cell Classification Using Fluorescence Imaging Multivariate Optical Computing.

摘要

Multivariate optical computing (MOC) is an all-optical approach of predictive spectroscopy that utilizes multivariate calibration and spectral pattern recognition techniques while operating in a simple filter photometer instrument, removing the need for expensive instrumentation and post-processing of spectral data. This is accomplished with specially designed interference filters called multivariate optical elements (MOEs).;MOC can provide analytical solutions for applications requiring low cost, rugged, and simple to operate instrumentation for use in remote and hazardous environments such as open ocean waters. These instrument specifications are central for developing a method for classifying phytoplankton in their natural environment. Phytoplankton are photosynthetic single cell algae and cyanobacteria that inhabit nearly all natural bodies of water The size and taxonomic composition of the phytoplankton community structure has global implications on carbon transport.;This dissertation describes the development of a single streak imaging multivariate optical computing (SSIMOC) method for single-cell classification of phytoplankton. The design and fabrication of MOEs for phytoplankton classification, along with an imaging photometer constructed for the purpose of collected data images for MOC analysis, will be discussed. Results of data collected with the SSIMOC on cultured phytoplankton and coastal water collected near the Martha's Vineyard Coastal Observatory will be shown.