INVESTIGADORES
KAMENETZKY Laura
artículos
Título:
A biologically-inspired validity measure for comparison of clustering methods over metabolic datasets
Autor/es:
STEGMAYER, G.; MILONE, D.; KAMENETZKY, L.; LÓPEZ, M.; CARRARI, F.
Revista:
IEEE-ACM TRANSACTIONS ON COMPUTATIONAL BIOLOGY AND BIOINFORMATICS
Editorial:
IEEE COMPUTER SOC
Referencias:
Año: 2012 p. 1 - 10
ISSN:
1545-5963
Resumen:
In the biological domain, clustering is implemented under the guilt by association principle, based on the assumption thatgenes or metabolites involved in a biological process are co-expressed/co-accumulated under the control of the same regulatorynetwork. Thus, a detailed inspection of the patterns grouped to verify their memberships to known metabolic pathways could bevery useful for the evaluation of the clusters from a biological perspective. The aim of this work is to propose a novel approach forcomparison of clustering methods over metabolic datasets, including in the validity measure the prior biological knowledge about therelation between elements that constitute the clusters. A way of measuring the biological significance of the clustering solutions isproposed, addressed from the perspective of the usefulness of the clusters to identify those patterns that change in coordination andbelong to common pathways of metabolic regulation. The measure resumes in a compact way the objective analysis of the clusteringmethods in which respects coherence and clusters distribution; plus it also evaluates their biological internal connections consideringcommon pathways. The proposed measure has been tested in two biological databases using three clustering methods.