The independent and identical distribution model successfully predicts CL distributions.

Quality of fit was assessed by squared-error between the fit and observed CL distributions (column 2) as well as the Pearson correlation r of the fit and observed CL data (column 3). 17/18 samples show r>0.7. The 4th column shows Pearson correlation of the fit and observed CL distributions in four-fold cross-validation (mean ± standard deviation). P-values compare the cross-validation correlation values to correlations calculated on CL data with the labels randomly permuted (1-tail t-test). All samples achieved p-value<0.05 indicating that the IID model could correctly describe the CL remodeling system.