"Frontmatter". In: Plant Genomics and Proteomics

C
OMPARATIVE
G
ENOMICS
The number of phylogenetic programs is continually increasing. One 
listing contains some 194 phylogeny packages (http://evolution. 
genetics.washington.edu/phylip/software.html), and no attempt will be
made to consider any of the merits or shortcomings of these programs here. 
The alignment of genomes across a wide range of species can add sub-
stantially to speculation concerning the function of various sequence regions
or motifs, and to the identification of genes. These alignments can be done
on the basis of sequence similarity, which can be scored on a nucleotide or
amino acid basis. An example would be the search for novel regulatory ele-
ments with phylogenetic footprinting (Blanchette et al., 2003). In this method
the assumption is that functional features evolve more slowly than non-
functional features. Where short motifs have been well conserved across evo-
lution, two different challenges arise. If the sequences are closely related, it
may be relatively easy to obtain a good multiple alignment. However, when
sequence divergence occurs, distinguishing between chance observation
versus selection pressure requires quantitative analysis. A statistical test that
uses known phylogenetic relationships between the species has been devel-
oped to determine whether the region is likely to be under selective pres-
sure (Blanchette et al., 2003). In one case, for closely related organisms, a
parsimony score is calculated and a P value associated with this is reported,
using a null model for neutral evolution. In the second case, when the 
promoter regions are quite diverged, a program called FootPrinter, which
identifies conserved motifs from a set of unaligned sequences, is allied to 
the phylogenetic information to evaluate a motif’s conservation. Results
obtained with a large genomic region in 13 vertebrates is shown on 
the genome browser at the University of California, Santa Cruz
(http://hgwdev.cse.ucsc.edu/zoo.html). The conserved regions that are
identified from these types of comparisons can then be subjected to experi-
mental verification. One of the specific advantages of using these compara-
tive approaches is the identification of important regions that lie far away
from the gene in whose control they are involved. A region that is approxi-
mately 60 kb 5¢ to the mouse Gdf6 gene (required for the development of
several skeletal joints) was shown to be critical for proximal limb joint
expression. This was identified with BAC transgene constructs and analysis
of transgenic mice. The identified region contained several evolutionarily
conserved noncoding sequences that are likely to be the core components 
of a limb joint regulatory element. However, there is evidence to suggest 
that additional regulation elements lie outside of the ~250-kb region that 
has been extensively characterized with the BAC transgenes. MultiPipMaker
(an informatic tool developed by Webb Miller, http://nog.cse.psu.edu/ 
pipmaker/) was used to compare those regions identified across a range of
vertebrates, resulting in the identification of a rodent-specific syntenic break
1 8 2
9. B
I O I N F O R M AT I C S

---

close to Gdf6, thereby making it unlikely that the additional regulatory ele-
ments could be identified by using sequence comparisons across a wide
range of mammals (Mortlock et al., 2003). 
As described in Chapter 1, syntenic relationships can be identified
among the flowering plants and beyond. It still awaits verification whether
the same type of long-range regulatory signals can be identified in plants as
have begun to be identified in animals. However, as the number of genomic
sequences increases, and the level of genomic sequence in a wider range of
plants becomes available, this type of analysis will become more feasible.
One of the advantages in animals is the availability of larger numbers of
“complete” genomic sequences. The proposals for plant genomic sequences
that are focused on the expressed or “genic” regions may fail to include those
regions responsible for any long-range regulation, so a case can be made for
including the “unimportant” repetitive regions in the sequencing of complex
plant genomes.
The genomic context of genes is also an essential piece of information to
help distinguish between paralogs and orthologs. The alignments that are
being developed and are available from the specific databases, such as
Gramene and MaizedB, are important resources in this task. The availabil-
ity of browsers that support a custom annotation track will allow individ-
ual users to build their own annotation structures, and these eventually will
substantially contribute to the distributed annotation effort.