Individual sequences in large sets of gene sequences may be distinguished efficiently by combinations of shared sub-sequences