Permutation testing for thick data when the number of variables is much greater than the sample size: recent developments and some recommendations