2.3 Classification Algorithm Methods

We have assembled algorithms(classification methods) in software that can automatically classify regions of interest in flow cytometry analysis. We will demonstrate that the particular populations required by our use cases can be validly, rigorously and repeatably identified automatically. In clinical tests, and diagnostic environments, automated region identification will eliminate a complex set of human instructions and decisions, reducing error and turnaround time.

Methods, results, and discussion are described in each of the individual reports.

Algorithm Result Overview
Algorithm Discussion Overview

Artificial Neural Networks - An ANN is a weighted directed graph where the nodes are artificial neurons, and weighted directed edges connect neuron outputs with neuron inputs. In this study, we considered both feed forward networks in which the graphs have no feedback loops (multilayer perceptrons and radial basis function networks) and feedback loop, or recurrent, networks (recurrent multilayer perceptron, competitive learning network, and Kohonen’s self-organizing maps).

Support Vector Machines(SVMs) are deterministic algorithms designed to identify the vectors within the training data that define the boundaries between classes of data.  Non-linear boundary problems are addressed using support vector machines by including a basis function that maps the input data into a transformed space that allows a linear discriminant to separate the classes.

Probability Bin Cluster Analysis (PBCA) - A method for population identification using Chi-squared analysis combined with an innovative binning strategy involving multidimensional spaces created by progressive histogram splitting to derive a model of multi-dimensional decision tree. We have commercial implementation of this published algorithm, and see promising qualitative results in the field, but have never conducted the detailed study to take the algorithms into an unsupervised environment.

Magnetic Gating – starting with a use case specific accepted position in the context of an assay, gates are applied in an iterative search pattern to find neighboring positions that maximize the evaluation function. Once a gate magnetic is made, the software moves it until it accommodates the maximum number of events--i.e., it relocates to where the events are. When a magnetic gate is copied from one sample to another, it is automatically readjusted. We intend to refine the evaluation functions to provide bias more towards the original position (anchoring) or to the position of a different gate (tethering), depending upon the application.