The Taxonomic Atlas of Prokaryotes
The atlas presents a series of interactive two dimensional maps of the
taxonomic space occupied by the Archaea and Bacteria and
is based on the analyses that were done in writing the revision to the
Road Map to the Manual of the . These maps provide a view of
the temporospatial relationships among the major lineages of cultured
and non-cultured species that are derived by performing a Principal Components
Analysis (PCA) on the covariance matrix computed from the evolutionary
distances among the 16S rDNA genes. PCA finds uncorrelated composite
measures in a multidimensional data set and allows expression of the
original data in far fewer dimensions, without a significant loss of
information. In the maps, each point represents the location of a given
sequence within the derived coordinate system, in reference to all other
samples within the data set. PCA has been widely used in ecology and
taxonomy and more recently has received considerable attention in the
genomics community as a tool for uncovering patterns of gene expression
in DNA microarray experiments.
For computational efficiency, we employ use an asymmetric matrix of
pairwise evolutionary distances as input for our analysis. Each sequence
(representing an individual species or strain) is treated as an independent
variable and its evolutionary distance to 223 reference sequences (which
we refer to as benchmarks) is computed. The benchmarks were selected
based on their designation as type or reference strains for the 184 families
in the described in the first revision of Bergey's
Taxonomic Outline of the Procaryotes or were reported as representatives
of major lineages of yet-uncultivated Bacteria or Archaeae. Using this
approach we are able to significantly reduce the dimensionality of the
original data to yield 2-D and 3-D views that account for >85% of
the total variance in the underlying data.
By linking evolutionary distance data to the hierarchical taxonomic
information in Bergey's Outline, it is possible to identify
the precise location of each sequence within PCA plots, as well as the
boundaries of higher taxa. This allows one to readily identify taxonomic
discontinuities and annotation errors and pose additional questions regarding
Page 1|2|3|4
| 
