It is critical to understand that this article fundamentally misuses the word "homology", which in bioinformatics and molecular evolution is understood to mean "sharing a common ancestor". Similarity searches (typically using programs like BLAST, one of the most cited methods in the biomedical literature) look for homologs.<p>This article is NOT about "homology" detection (which suffix trees are not particularly good for), it is about "identity" detection (which is a related problem, but fundamentally different). It is unfortunate that in the biomedical literature, the word "homology" is often misused in place of "identity" (as in "micro-homology", which is really "micro-identity"), but serious computational biologists try to avoid this misuse of the word "homology".
There is a rich literature in bioinformatics on sequence homology search, and there are many existing libraries that scale to billions on base pairs. I wonder why they reinvented the wheel