One embodiment provides for a method including determining, by at least one processor, sequence-sequence distances for a biological sequence collection. The at least one processor generates a matrix Mij of the sequence-sequence distances, where i and j are positive integers. The at least one processor further generates clusters for the matrix Mij by performing hierarchical clustering. A self-consistent taxonomy is created from the clusters. A visual heat map display of the matrix Mij is selectively controlled using metadata, zoom input and opacity input.