Trends in Biotechnology
Vol.19 No.1 January 2001
http://tibtech.trends.com 0167-7799/01/$ – see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S0167-7799(00)01519-5
6
News&Comment
Journal club
Cataloguing life
At my last count, and if you include
organelles and viruses, more than 700
complete genomes have been sequenced so
far. Bacterial genomes make up the major
portion of the organisms sequenced; 32 have
been completed and 107 are in progress.
Only three eukaryotic genomes, those of
Caenorhabditis elegans
,
Saccharomyces
cerevisiae
and
Drosophila melanogaster
have been completed (see the websites:
http://www.tigr.org/tdb/mdb/ mdbcomplete.
html; http://www.ebi.ac.uk/ genomes/
info.html). In addition, a draft of the human
genome and a large number of human
expressed sequence tag (EST) sequences are
available and a draft of the mouse genome
should be finished shortly. It has become
routine to generate catalogues of genes and
the next phase of understanding the content
of the catalogues is just beginning.
Why is the sequencing of the
Pseudomonas aeruginosa
genome,
seemingly just another bacterial genome
important? The genus
Pseudomonas
is not
only diverse but is essentially made up of
free-living organisms. The majority of
bacterial species of which the genomes have
been sequenced to date have a parasitic or
symbiotic lifestyle, which means that they
live in a relatively defined environment.
Pseudomonads, by contrast, are just as likely
to be found in the soil, the oceans, and in
contaminated food as in the lungs and
wounds of immunocompromised patients.
The striking feature of the
P.aeruginosa
genome is that its complexity in terms of
potential gene products approaches that of
the simple eukaryote
Saccaromyces
cerevisiae
. However, comparison of these
two genomes based on our current
understanding showed very little direct
homology between these two organisms.
The highest level of similarity that was found
by the authors was between
P. aeruginosa
and
Escherischia coli
, another well studied
gram negative bacterium. Even in this case
merely 48% of the genes gave significant
matches, with a median amino acid identity
of merely 40%. The work published in this
paper
1
should lead to a better understanding
of the types of functional diversity that are
required by a free-living bacterium to survive
and successfully compete with other species.
P. aeruginosa
is also notable for its antibiotic
and disinfectant resistance. These abilities
are reflected in the complexity of its genome
and the products it can generate. Detailed
studies of its products will lead to a better
understanding of antibiotic resistance in this
species and probably in other species as
well.
The many genome sequencing projects,
including the human genome project, are
providing a vast amount of information akin
to the hieroglyphics on ancient Egyptian
tombs. The challenge now facing biological
research is to find the Rosetta stone(s) to
enable us to decipher the text and translate it
into functional relationships between the
many individual proteins and sets of proteins
that make up a living organism. Comparing
the genome databases can yield important
clues about the differences between species
and act as a guide as we try to fill the many
accumulating genetic catalogues with
detailed product descriptions.
Keith Ashman
Ashman@mshri.on.ca
1 Stover, C.K. et al.(2000) Complete genome
sequence of Pseudomonas aeruginosaPAO1, an
opportunistic pathogen. Nature406, 959–964
PII: S0167-7799(00)01516-X
Metabolic networks: the bigger picture
The avalanche of data that genomic
sciences are continuously producing has
confronted biologists with the daunting task
of making sense of it. Lacking a conceptual
understanding of the fundamental
principles of the living organism, we have
resorted to brute-force statistics to cluster
and classify molecules and to correlate
patterns of expression profiles. However,
these current genocentric approaches fail
to embrace the genome as a whole, and
address basic principles of its organization
as a functional entity. The latter will be
important because our knowledge of the
molecular machinery of the cell has
accumulated over the past decade, and
thus the picture of simple linear gene-to-
function pathways is being replaced by the
notion of one genome-wide biochemical
network. But what is the very nature of this
network? What is its generic ‘wiring
architecture’?
Now, a group of physicists around Albert-
László Barabási at the University of Notre-
Dame, together with pathologist Zoltan Oltvai
at Northwestern University, report in
Nature
their work on precisely these generic features
of biological networks, taking advantage of
the availability of metabolic pathway
databases
1
. Barabasi
et al.
have previously
studied the architecture of complex networks,
such as the Internet or social networks, and
found that such evolved networks share
fundamental generic properties. In general
terms, a network consists of nodes connected
by edges. By analysing the large-scale
structure of such networks, they found that
they belong to the class of ‘scale-free’
networks of which the wiring architecture
deviates significantly from the random,
‘exponential’ networks. Scale-free networks
are highly inhomogeneous: they contain
nodes that have a significantly higher number
of connections than the average, acting like
hubs. Scale-free networks are robust to errors:
random deletion of individual nodes does not
significantly affect information flow in the
network. However, they are vulnerable to the
removal of a highly connected node. The
group now extends their analyses to
metabolic networks of 43 organisms, which
include Archaea, Bacteria and Eukarya, in
which the substrates are the nodes and the
metabolic reactions are the connecting edges.
They show that the metabolic networks of all
the examples studied display the topological
properties of scale-free networks: they are
dominated by a few highly connected
substrates that turned out to be the same
across the species. The scale-free property
indicates that the architecture of metabolic
pathways is organized such that metabolism
is robust and error-tolerant but is also
sensitive to attacks on the highly connected
nodes, which might explain the effectiveness
of drugs. In addition, the metabolic networks