New TreeView Display Option in NCBI BLAST

The new Tree View option on the NCBI Web BLAST service presents a dendrogram or tree display that clusters sequences according to their distances from the query sequence. This display is helpful for recognizing the presence of aberrant or unusual sequences or potentially natural groupings of related sequences such as members of a gene families or homologs from other species in the BLAST output.

The link labeled 'Distance tree of results' that leads to the Tree View display appears on the BLAST output for all DNA-DNA or protein-protein comparisons. Trees can be rendered as rectangular, slanted (cladogram), radial or force displays by the selecting the corresponding tab from the tree output. Both the rectangular and radial outputs are scaled to show the distances between sequences. Subsets of the tree or any of the alignments can be displayed through the 'Show subtree' or 'Show alignment' links pop-up menu that appears at the internal nodes of the tree on mouse-over. The following two examples show how the tree display can produce useful trees for nucleotide or protein sequences.

A Tree based on nucleotide-level comparisons

Fig. 1 shows a radial tree display generated by searching against the 'refseq_genomic' database with the woolly mammoth complete mitochondrial genome (RefSeq accession NC_007596). The RefSeq genomic database was limited to the mammalian taxon 'afrotheria'. This tree reconstructs the accepted taxonomic groupings of these mammals and reinforces the proposition that the woolly mammoth is most closely related to the African and Asian elephants.

Click on image to view larger

Figure 1. Radial tree display generated by searching against the refseq_genomic database with the woolly mammoth complete mitochondrial genome (RefSeq accession NC_007596, shaded). The mammoth sequence, highlighted, clusters closely with that of the modern elephants.

A protein tree is shown in Fig. 2. This tree was generated from the results of a search using a Trypanosome arginine kinase (RefSeq accession XP_826998) against the Swiss-Prot database. This tree shows two distinct groups of related enzymes in the results, creatine kinases from vertebrates and arginine kinases from arthropods, a few other invertebrates, and the trypanosomes. The tree highlights the surprisingly close relationship between the trypanosome proteins and the arthropod proteins that has given rise to the hypothesis that the trypanosome arginine kinase genes were acquired by horizontal transfer from an arthropod host of these parasites.[1]

Click here for live results

Figure 2. A portion of the rectangular tree generated from a search using a Trypanosome arginine kinase (RefSeq accession XP_826998) against the Swiss-Prot database showing the invertebrate arginine kinases. The terminal nodes are labeled by 'blast name' to highlight taxonomic trends. The trypanosome query sequence clusters with the arthropod sequences suggesting the possibility of horizontal gene transfer between these phylogenetically distant groups. A sequence from the flatworm Schistosoma mansoni clusters with four molluscan sequences suggesting another possible case of gene transfer.

Method and caveats

The BLAST tree display is created from genetic distances calculated using standard methods from the aligned sequences— Jukes-Cantor[2] for nucleotide comparisons, Kimura’s method[3] for proteins. The trees themselves are then built from these distance matrices using either the Fast Minimum Evolution (FastME)[4], or Neighbor Joining[5] methods. Since BLAST is used to create the alignments, the database sequences are only compared and aligned to the query, not to each other as they would be in a multiple sequence alignment. Therefore it is important to keep in mind that the alignments used to create the Tree View display may differ from the global multiple sequence alignments typically used to infer phylogenies. But, as the above examples show, with judicious choice of query sequence and control over the database sequences searched very informative trees can be generated.

—TT

[1]Pereira CA, Alonso GD, Paveto MC, Iribarren A, Cabanas ML, Torres HN, Flawia MM. Trypanosoma cruzi arginine kinase characterization and cloning. A novel energetic pathway in protozoan parasites. J Biol Chem. 2000 Jan 14;275(2):1495-501. PMID: 10625703
[2]T. Jukes, C. Cantor, in Mammalian Protein Metabolism, H.N. Munro, J. Allison, Eds. (Academic Press, New York, 1969, vol. 3 pp. 21-32).

[3]M Kimura. The Neutral Theory of Molecular Evolution. Cambridge University Press, 1983.

[4]Desper R, Gascuel O. Fast and accurate phylogeny reconstruction algorithms based on the minimum-evolution principle. J Comput Biol. 2002;9(5):687-705. PMID: 12487758.

[5]Saitou N, Nei M. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol. 1987 Jul;4(4):406-25. PMID: 3447015.