LumaCyte’s Radiance instrument measures dozens of parameters that reflect the intrinsic biophysical and biochemical properties of cells. Multivariate data (data consisting of multiple variables as opposed to a single variable) can have significant benefits for qualitative and quantitative analytical modeling/calibration, and better predictive results versus simple univariate data sources. The measurements can be used to build a sophisticated model that can be used to better understand a process, identify new patterns, and predict important variables more precisely and accurately.
To the left is a selection of data collected with Radiance from a 3-component particle mixture including eccentricity (non-spherical aspect ratio), deformability (particle stretch while experiencing optical forces), and velocity (proportional to optical force). The video shows that the particles are partially resolved in eccentricity and deformability but have significant separation in the velocity dimension.
The power of a multivariate analysis tool is that even for very overlapped data, small differences can be utilized and correlations established using a vast array of machine learning tools including partial least squares regression (PLS) analysis, principal component analysis (PCA), linear discriminant analysis (LDA), and neural networks.