Methods for Analyzing High Dimensional Data for Classifying, Diagnosing, Prognosticating, and/or Predicting Diseases and Other Biological States

This invention relates to a method of using supervised pattern recognition methods to classifying, diagnosing, predicting, or prognosticating various diseases. The method includes obtaining high dimensional experimental data, such as gene expression profiling data, filtering the data, reducing the dimensionality of the data through use of one or more methods, training a supervised pattern recognition method, ranking individual data points from the data, choosing multiple data points from the data based on the relative ranking, and using the multiple data points to determine if an unknown set of experimental data indicates a diseased condition, a predilection for a diseased condition, or a prognosis about a diseased condition.

Artificial neural networks (ANNs) are computer-based algorithms capable of pattern recognition particularly suited to making diagnoses. ANNs do not require explicit encoding of process knowledge in a set of rules and can be trained from examples to recognize and categorize complex patterns. ANNs learn more efficiently when the data to be input into the neural network is preprocessed. Various ANN approaches to the analysis of data have seen extensive application to biomedical problems, including those in the areas of diagnosis and drug development. Unsupervised neural networks are also extensively used for the analysis of DNA microarray data.