Biology Reference
In-Depth Information
at least was not. In the pre-genome sequence era, the
approach to identifying the causative mutation was limited
by sequencing power, and proceeding in broad terms as
follows (reviewed by Fay et al. in depth in WormBook
[44]
). During a screen multiple mutants are typically iso-
lated; for each mutant, the locus responsible for the mutant
phenotype of interest is first crudely mapped to linkage
groups, then to large regions of a specific chromosome
using classical genetics. Guided by this crude mapping,
complementation tests are carried out to identify how
many independent loci have been isolated in the screen. For
any single locus, the next step is to rescue the mutant
phenotype
now do rapid SNP mapping of the locus to a handful of
genes, a huge advance. The other critical advance is in
sequencing power
using 'next-generation' sequencing
technology of any type currently gives sufficient coverage
of the 97 Mb worm genome to confidently identify any
single-base mutation for around $100. Combining SNP-
mapping with whole-genome sequencing therefore allows
the extremely rapid identification of causative mutations
[46,47]
, and this should open the way to deeper, more
saturating forward genetics. This in turn should provide
more accurate frameworks for the systems analysis of
pathways leading to quantitative and predictive modeling,
as has been attempted for the vulva (for example, by
[48
e
whether as cosmids, YACs, fosmids or BACs,
some large fragments of the genome covering the region
identified by genetic mapping are injected into the mutant
animal and the resultant transgenics tested for rescue. This
is repeated using increasingly smaller test fragments until
a fragment is isolated that is small enough to sequence
using standard Sanger sequencing and traditionally,
together with cDNA library probing and Northern blots the
transcriptional unit that rescues the mutant phenotype can
thus be identified. Finally, having identified the gene
affected, PCR and sequencing can home in on the sequence
changes in the mutant, and the story is therefore complete,
one gene at a time.
The process outlined above is clearly laborious and
extremely low throughput. Cloning a single gene, along
with some basic molecular characterization, was typically
the body of a 5-year PhD. As a result, there is often
extensive triaging of the mutants to be mapped
e
51]
) and which one hopes might one day approach the
beautiful fusion of theoretical and experimental analysis of
signaling in bacterial chemotaxis (reviewed in
[52]
).
e
Persuading Worms to take Enough Drugs
e
Chemical Genomics in the Worm
Although most of this section focuses on using genetic
perturbations
phenotype
connection (whether through mutagenesis or using RNA-
mediated interference), chemical genomics provides
a completely different entry point to perturbation of genetic
networks (reviewed, for example in
[53]
). Chemical
screens can not only identify vital medical compounds, but
on a research level provide many crucial tools
to
examine
the
genotype
e
it is hard to
imagine mammalian signaling research without the sets of
compounds used to activate or repress specific pathways.
However, chemical genomics in the worm is complicated
by a specific feature of normal worm biology: worms live in
dirt. To deal with this complex environment, they have
evolved a defensive arsenal of cytochrome P450s and drug
transporters to ensure that any potentially harmful mole-
cules to which they are exposed are either rapidly modified
or removed from their bodies
[54]
. In the standard labora-
tory environment this makes little difference: unchallenged
by toxins, these defensive weapons lie unused. However,
this raises clear difficulties for chemical genomics, as many
of the chemicals in commercially available compound
libraries are either rapidly modified, rendering them
broadly ineffective, or pumped out of the worm so that
insufficient intracellular concentrations build up and no
effect is seen. The cuticle is also a major barrier to drug
entry, and while some drugs penetrate relatively easily,
others must be applied to the intact animal at hundreds to
thousands times higher concentrations than their estimated
target affinity, increasing cost and, in some cases, leading to
solubility problems. Together, these issues have limited the
number of de novo compound screens carried out in the
worm (compared, say, with the extensive small molecule
screens carried out either in yeast or in mammalian cells in
e
mutants
with low-penetrance phenotypes are often frozen away and
neglected, and anything with complex phenotypes indi-
cating extensive pleiotropy is often ignored. Researchers
understandably initially focus on the clean and the strong,
100% penetrant phenotypes with no confounding addi-
tional defects. If one can only identify one or two genes
a year, this makes sense, but it does lead to an ascertain-
ment bias in the hits from these screens that is hard to
quantify and hard to compensate for in network or systems-
level analysis of the molecular pathways identified.
Why was the throughput of these screens so low? The
problem was not usually that of screening enough worms,
or picking enough mutants, but rather the slow process of
mapping leading to the isolation of a small enough piece of
rescuing DNA that was worth sequencing. The availability
of the genome sequence and the huge recent advances in
sequencing power have changed this landscape massively,
however. For mapping, the identification of many SNPs
between the canonical N2 Bristol isolate sequence
[26]
and
the CB4856 (often called the Hawaiian isolate) has revo-
lutionized the mapping of mutant loci in the worm
[45]
.
Rather than rely on visually obvious mapping markers,
which are only dense enough to typically map down
a mutant locus to a region of around 100 genes, one can
e
