When I ran my PCR 16S V 9 -DGGE on environmental water samples from the Rio Grande, I detected the bands using Quantity One by BioRad. Unfortunately, I ran duplicates on the gel. For each of the three sites, 28 bands were detected and unknowns matched as many as I could, not all bands on the lanes could match be matched. For the similarity matrix we used the cluster method based on band position by lane only. So when I ran the UPGMA, one of the sites separated between the other two sites. So, I am wondering why would it do that? The other sites remained together except that site, this are duplicates!