Check homogeneity of selected dastasets in h5_expression_data

[lucavd]

A function that tests if the selected datasets are different while the number of datasets gets higher.

Difference to get tested as SD? Or what else?