Peter Hallin, Tim Binnewies and David Ussery at the Technical University of Denmark at Lyngby examine how the BLASTatlas tool can be used to spot the differences between similar genomes

One of the central tenants of molecular biology is that biological information flows from DNA to RNA to protein. Furthermore, a protein's sequence determines its structure and the structure its function. In general, although there are many methods that attempt to predict protein structure, it is still often difficult to predict the structures of novel sequences. Currently the state of affairs is pretty much limited to homology searches - that is, if one can find a good sequence match to a protein with known structure and function, it is inferred that the unknown protein has the same function. One can question the reliability and wisdom of this, but nonetheless, this is a commonly used approach.

There has been a literal explosion at the level of sequence information. The development of fast and inexpensive methods for sequencing bacterial genomes has led to a wealth of data, often with many genomes being sequenced of the same species or of closely related organisms. Thus, there is a need for simple ways to compare these sequenced genomes. We have developed the BLASTatlas method for mapping and visualising whole genome homology of genes and proteins within a reference strain compared to one or more other strains or species of prokaryotic organisms. The map is accurate at the level of amino acid conservation - in many cases few regions of a protein will be conserved.

Zooming in: a BLASTatlas map allows genomes to be compared

The figure shows a zoom of a particular region of a BLASTatlas for the main chromosome of Clostridium botulinum strain Langeland A, compared to 11 other Clostridia genomes. There are three dark blue lanes near the top, where most of the proteins are conserved, implying that these proteins probably play the same functional role in these organisms. The other genomes compared show little sequence conservation overall, and the genes contain only partial matches. So the plot answers questions not just about the presence or absence of genes, but about the degree to which they are conserved. The strong sequence conservation in the three dark blue lanes makes biological sense in that they all represent different strains of C. botulinum. The diversity within the genus of Clostridium is amongst the largest of all bacterial genera, and is as great as that between all animals.

Gene regulation is controlled in part by the mechanical-chemical properties of chromosomal DNA. So we map DNA structural properties from the reference genome sequence on the same atlas, showing amino acid conservation of protein coding genes. The bottom three lanes in the figure represent DNA structural parameters that we routinely use. Stacking energy measures how easily the strands in the DNA helix separate, position preference is a measure of the DNA's anisotropic flexibility and curvature reflects the three-dimensional structure of a short region of the DNA helix. Extreme values before the start of a gene indicate the potential presence of a promoter - a region of DNA that precedes a sequence to be transcribed into RNA. At the chromosomal level, sometimes these structural parameters correlate with functional properties, such as highly variable regions, or regions containing highly expressed genes.

We allow users to directly access the tools that are used to generate these BLASTatlas maps, through the use of Simple Object Access Protocol, commonly known as SOAP. This framework allows clients to directly communicate with our server and in a programmatic fashion, thereby allowing the user's own programs to systematically examine a large number of genomes. 

Read Hallin  et al's highlight in a forthcoming issue of Molecular BioSystems.