THE SIMCAL FAMILY OF ALGORITHMS FOR ANALYSIS OF MICROARRAY DATA

SiMCAL 1, 2, and 3 (Simple Multilevel C lustering And Linking, versions 1, 2, and 3) are novel clustering algorithms for the analysis of microarray data. The purpose of these algorithms is to present complete feature sets not found in either Jarvis-Patrick clustering, from which the original SiMCAL concept is derived, or in other popular clustering methods such as hierarchical and k-means. Although each algorithm in the SiMCAL family has distinct features and methods, they all share the following attributes: they are simple, in that they are computationally inexpensive; they are multilevel, in that they provide a small number of clearly defined hierarchical levels of clusters; and they offer linking between clusters at the same level in each hierarchy. Presented here are the design, development, and analysis of the algorithms; their applications to two types of microarray data, one involving the phosphatidylserine receptor (PSR) and the other involving cystic fibrosis (CF); a description of the Web-based interface for visualization of results; and possible avenues for further development. Code and data are available at http://www.dvorkin.com/daniel/Simcal123.zip under an open-source license.