Parallelisation using make_peak_matrix

[newtonharry]

I noticed this function doesn't take n_jobs unlike add_tile_matrix. Does this mean it is not parallelised or does it internally use a default number of threads/cores?

Also as a suggestion, perhaps rename this to add_variable_tile_matrix since you can specify custom genomic ranges regardless if its actually a peak or not.