This is a summary of MATLAB tools I developed to facilitate PCA analysis
---Mahalanobis-Distance and getCovMatrices----
Main (Mahalanobis-Distance):
This is a tool to determine if there is a statistical difference between two subgroups in a PC1-PC2 coordinates system
It follows the routine demonstrated in https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4523310/
It includes the calculation of Mahalanobis Distance followed by F-test statistics
The program is designed for 2 variants (herein PC1 and PC2) and 2 subgroups (for example treatment and control group)
Input here is an Excel Table with following format
Columns: VAR1_group1 - VAR2_group1 - VAR1-group2 - VAR2_group2
(2nd col) (3. col) (4. col) (5. col)
herein VAR1 = PC1
VAR2 = PC2
User input: change in the code of MahalanobisDistance (main routine) the name of the sheet and insert number of groups (it's 2 as default, I recommend to leave that)
Output: DW = Mahalanobis Distance
Tsqr = two sample Tsquared
F = F-Value
Function getCovMatrices is called to calculate the pooled between-group covariance matrix (according to https://blogs.sas.com/content/iml/2020/07/01/pooled-covariance-between-group.html)