Principal component and cluster analysis of morphological variables reveals multiple discrete sub-phenotypes in...

The present study evaluates the usefulness of the principal component analysis-based cluster analysis in the categorization of several sub-phenotypes in the weaver mutant by using several morphological parameters from the cerebellar cortex of control, heterozygous (+/wv) and homozygous (wv/wv) weaver mice. The quantified parameters were length of the cerebellar cortex, area of the external granular layer, area of the molecular layer, number of the external granular layer cells (EGL), and number of Purkinje cells (PCs). The analysis indicated that at postnatal day 8, the genotype +/wv presented three sub-phenotypes tagged as +/wv (0), +/wv (1) and +/wv (2), whereas two sub-phenotypes designated as wv (0)/wv (1) and wv (0)/wv (2) were identified in the genotype wv/wv. The number of PCs for the genotype +/wv and the number of EGL cells for the genotype wv/wv were the variables that discriminated the best among sub-phenotypes. Each one of the sub-phenotypes showed specific abnormalities in the cytoarchitecture of the cerebellar cortex as well as in the foliar pattern. In particular, the wv (0)/wv (1) and wv (0)/wv (2) sub-phenotypes had the most altered cytoarchitectonics, followed by the +/wv (2) sub-phenotype and then by the +/wv (1) one. The sub-phenotype +/wv (0) was the less affected one. Apart from reporting for the first time the coexistence of several sub-phenotypes in the weaver mutant, our approach provides a new statistical tool that can be used to assess cerebellar morphology.