A phylogeny based on one metabolic gene conflicts with an rRNA-based phylogeny, possibly because the metabolic gene was horizontally transferred. Why might the rRNA tree still be informative for deep ancestry?