Characteristics and value of directed algorithms in high content screening

Methods Mol Biol. 2007:356:63-81. doi: 10.1385/1-59745-217-3:63.

Abstract

High content screening requires image processing algorithms that can accurately and robustly analyze large image numbers without requiring human intervention. Thus, a suite of algorithms that are directed by an understanding of the biology being studied was developed for the optimized automated acquisition and quantitation of cellular images. Two categories of directed algorithms were developed: Developer Tools for assay development and Specific Algorithms for turnkey screening of specific biological situations. The same basic sequence of analysis steps are used in these directed algorithms: 1. Primary object identification. 2. Measurement of primary object properties. 3. Identification and measurements of associated targets. 4. Analysis of raw measurements for specific biological problems. The detailed application of these steps is guided by the biology being studied and the expected phenotypic changes. Most cell biological problems to be analyzed using high content screening can be categorized by either the phenotype of the problem or labeling pattern, or by a standard biological response behavior of the cells. This enables application of directed algorithms optimized for these categories. Examples of the use of directed algorithms for specific categories are discussed, as well as the detailed analysis steps for a specific directed algorithm.

MeSH terms

  • Algorithms*
  • Animals
  • Image Processing, Computer-Assisted
  • Software
  • Tissue Array Analysis / methods*