biostatistics - sampling size

Hi. I am interested in discovering motifs in a set of DNA sequence(say 100) and I would like to use the approach of using a subgroup(say 10 from 100) and likewise create random subgroups(picking 10 from 100 randomly) to discover the motifs in the given set of DNA sequences. I would like to know from statistics point of view, how many number of genes should be in a single subgroup and how many total subgroups would be required for this analysis.