In my experience with 16s, 23s and other ribosomal RNA sequence alignments, the gappy regions corresponding to loops or less structured regions of the sequences, are not phylogenetically informative across large distances (such as all vertebrates, or all insects) but can be useful over shorter distances (such as comparing feline mammals or one lineage of ants or bees. The problem is that insertions and deletions happen such that the sequences in these regions can have similar length but not be homologous (shared a common ancestor) because they represent different insertion or deletion events.

Phlylogenies built from single genes are most often not truly representative of the complete genomes, anyway. They can be very useful for many things, but it is not realistic to expect any one gene to give us a true picture of the evolution of any group of species. Phylogenetic trees built from bacterial 16S rRNA for example are often good enough to place each sequence in the "ballpark" such as "Enterobacteria" but not good enough to tell us exactly which species of bacteria the 16S sequence came from.