High-content screening (HCS) via high-resolution fluorescence cellular microscopy is a powerful technology for drug discovery and is currently being used in secondary screening and optimization of candidate compounds.
The aim of Cytoprint's proposed effort is to demonstrate advanced image analysis informatics tools to address critical bottlenecks in the deployment of HCS technology, which are the limited availability of assays and the lack of a reference database and tools for indexing, cataloging and searching cellular imagery data bases. Cytoprint converts cellular imagery into a numerical signature that can be ordered and indexed using an information-theoretic similarity metric. The numerical signature is obtained via concatenation of the results of several different image feature extraction operators applied to the imagery. Cytoprint has successfully demonstrated the ability of this approach to detect and differentiate between apoptotic and non-apoptotic compounds based upon signatures extracted from cellular imagery. The objective of our Phase I SBIR effort will be to refine and verify this analysis on a larger image set. The objective of a subsequent Phase II SBIR effort will be to expand the method to enable differentiation among several distinct biological effects.
