Scientists at Applied Maths publish novel strategies to assess and improve the reliability of cluster analysis

Scientists working at Applied Maths have published a research article that describes a global statistical framework to calculate the reliability of branches on trees or networks resulting from cluster analysis, and that allows the most probable trees to be produced from a set of degenerated solutions. The article, titled “A resampling strategy for reliable network construction”, has appeared in the May 2011 issue of Molecular Phylogenetics and Evolution (MPE 60: 273-286).

Cluster analysis of biological data, interpreting trees and assessing the reliability of branches has always been a central theme in the development of software at Applied Maths. The success of its software can be attributed for a large part to the availability of powerful clustering applications that are user-friendly, offer rich visualization and interpretation environments and allow clustering of the largest and most complex biological data sets. With the present article, describing algorithms that are implemented in the BioNumerics software, Applied Maths confirms its position as a pioneering and leading provider of powerful and user-friendly software for mining and clustering of biological data.

Luc Vauterin, Director of Science at Applied Maths: “With the emergence of new high-throughput techniques, turning data into information becomes a bottleneck, in spite of ever increasing computing power. Cluster analysis helps finding signals and structure in complex data sets of whatever source and nature. Clustering techniques have become increasingly popular in all branches of bioinformatics, ranging from phylogeny to drug discovery. Applied Maths’ mission is to provide scientists with powerful clustering tools that support reliable interpretation of the result.”

The full article can be downloaded from ScienceDirect.