Microglial Morphology Analysis via PCA

Principal component analysis (PCA) was performed in the RStudio Software to further evaluate morphological clustering of microglia using the 13 remaining measured parameters (excluding those used for LDA): NOB, FD, LAC, CP, CHSR, MSACH, CHA, CHP, R, CHC, TRMM, MR and DOB. To select the Principal Components (PCs) that represented the systematic sources of variation in our data and discard PCs that only reflect random noise, a permutation-based test was employed [78 –80 (link)]. PCAtest (https://github.com/arleyc/PCAtest) was used to evaluate the significance of each PC and of the variable loading for the significant axis [80 (link)]. This function applies a permutation-based test and builds a null distribution to be compared for each parameter [80 (link)]. The cells grouped in the four ROIs identified with the HCA were included in this analysis (243 microglial cells).