MULTIFACTORIAL COMPARATIVE-STUDY OF SPATIAL POINT PATTERN-ANALYSIS METHODS

A way of studying cooperative behaviour of biological entities (protei ns, cells, etc.) is by using topographical analysis: the quantificatio n of the spatial patterns formed by the entities considered as points. Five methods of topographical analysis were compared in terms of disc riminant power, stability of parameters, methodological bias and algor ithms. We tested five methods (nearest neighbour distribution, radial distribution, Voronoi paving, quadrat count, minimal spanning tree gra ph) which generated nine parameters on four simulated models (random p oint process, hardcore model and two cluster models) and on experiment al cellular models. The method which offers the best discrimination po wer and stability seems to be the minimal spanning tree graph edge len gth distribution. (C) 1997 Academic Press Limited.