Cluster Analysis Method in Electrophoretic Patterns Classification

F. Spirovski^a,*, K. Stojanoski^a and A. Mitrevski^b 

^aInstitute of Chemistry, Faculty of Natural Science, Sts.Cyril and Methodius University, 1000 Skopje, Macedonia
^bClinic of Neurology, Faculty of Medicine, Sts.Cyril and Methodius University,1000 Skopje, Macedonia

  The conventional electrophoresis methods are well known techniques for protein detection and analysis of cerebrospinal fluid (CSF). Disc electrophoresis (DEP) was carried out for detection of oligoclonal IgG bands in cerebrospinal fluid (CSF) on polyacrilamide gel. However, the advance of automation has made rapid collection of large amounts of data feasible and the development of microcomputers has made sophisticated processing even of old electrophoregrams possible. Automated analysis, data storage and sophisticate data acquisition were carried out with Gel Pro Analyzer 3.1, which is specifically structured to analyze gels and elctrophoregrams: complex band pattern matching (gel variation, dendogram analysis etc.); lane relation studies (sophisticated lane database); general gel analysis (accurate molecular size, quantitative determination of protein mixture etc.). Clustering techniques have been applied for detection of intrathecal immune response. Different hierarchic cluster analysis methods such as single linkage, complete linkage, unweighted pair-group average (UPMGA) were used. In addition, other cluster characteristics such, distance matrix and Euclidean distance matrix were calculated. Pairing of electrophoresis methods and cluster image analysis, could lead to additional diagnostic information of inflammatory conditions of the central nervous system (CNS) or dysfunction of blood-CSF barrier.

Keywords: Disc electrophoresis, Cerebrospinal fluid, Immunoglobulin G, Cluster analysis