Biology Reference
In-Depth Information
phenotype connection that
cannot be detected trivially using forward genetics screens
is the early developmental role for genes that have a zygotic
requirement but which also have a large maternal contri-
bution. The early events of embryogenesis are almost
completely reliant on the large amounts of mRNA and
protein contributed by the maternal germline, and so
embryos that are homozygous null for a gene that plays an
essential role in early embryogenesis may nonetheless go
through early embryogenesis absolutely normally if it was
laid by a heterozygous mother. Since many of the genes that
play key roles in early embryogenesis also function later in
the life of the animal, one cannot derive embryos from
homozygous null mothers (and thus eliminate maternal
contribution). The roles that such genes play in the key
initial events of embryogenesis are thus hidden from clas-
sical genetics, and developing models of the machineries
underpinning embryogenesis
Another class of genotype
specificity of such interactions from classical screens. For
example, mutations in hmg-1.2, a HMG-box transcriptional
regulator, strongly enhance the effect of mutations in the
Wnt pathway on male ray development
[38]
. The conclu-
sion drawn could therefore be that hmg-1.2 is a specific
regulator of Wnt signaling in male ray development.
However, systematic studies show that reductions in hmg-
1.2 activity enhance mutations in a wide variety of
signaling pathways, including EGF, Wnt, Notch and Ephrin
pathways
[39]
. Therefore, just as testing a single pairwise
protein
e
protein interaction can give a positive result that
may be misleading owing to the 'stickiness' of one of the
proteins, so a genetic screen to identify enhancers can be
misinterpreted due to the promiscuity of the enhancer and
the lack of systematic pairwise testing. This creates prob-
lems for any downstream systems-level analysis, as without
a measure of specificity the importance of the involvement
of a gene in a given pathway may be overestimated.
Despite the caveats of classical screening in the worm,
the components identified in such screens and their
ordering in pathways through epistasis analysis have yiel-
ded key frameworks that served to direct the hypothesis-
driven detailed molecular analyses that ultimately fleshed
out the physical basis for the genetic pathways. For
example, in the cell death pathway, genetic screens iden-
tified three loci that were critical for programmed cell
death, ced-3, ced-4 and ced-9; ced-9 was then shown to be
genetically upstream of ced-3 and ced-4. Following the
molecular cloning of these genes, and through biochemical
and molecular analyses by several groups, we now know
that in the 959 cells that survive, CED-9 physically inter-
acts with CED-4, restraining it from binding the pro-
apoptotic form of CED-3
[40
e
is
therefore
impossible
without alternative approaches.
A third issue, particularly from the point of view of
systems biology, is the difficulty of estimating how near
any mutagenesis screen is to saturation. Although approx-
imate calculations can be done to estimate the coverage of
a mutagenesis screen, it is clear that mutagenesis is far from
random, and that even screens which isolated multiple
independent alleles of the same gene that have the exact
same mutation (e.g., more than one allele of the G13E
mutation of the ras orthologue let-60 were identified in
genetic screens for vulval mutants
[35]
), which one might
therefore imagine to have reached saturation, missed
multiple genes that were subsequently identified to affect
the same process. This raises clear issues: if it is hard to
estimate coverage and false negative rates of any screen, it
makes systems-level analysis of any identified pathway
extremely precarious. In Rumsfeld-speak, while 'known
unknowns' (measurable false negatives rates) are tolerable
for systems biology, the 'unknown unknowns' are not. The
under-saturation of many screens is obvious
42]
; release of CED-9 from
CED-4 in the 131 cells destined for death, largely through
interaction with the BH3-only containing protein EGL-1
[43]
, releases CED-4 and it binds CED-3, leading to CED-3
auto-proteolysis to yield the mature active cysteine
protease which is the key apoptotic effector. This core
apoptotic module is highly conserved, and classical screens
were critical for the identification of all the components,
which we now understand at atomic resolution. Classical
genetics has been crucial in defining similar core molecular
modules for processes ranging from mechanosensation to
sex determination, and these genetically identified modules
can serve as the starting point for rigorous modeling
approaches such as those being pioneered in the vulva.
e
many genes
that were isolated through forward genetics screens have
only had a single allele isolated, suggesting that there must
be many potential 'hits' that fell just below this threshold
and thus escaped identification.
Finally, many genes have weak loss-of-function
phenotypes that may be hard to detect in an otherwise wild-
type background; for this reason, modifier screens have
often been used to identify additional components of
specific pathways that eluded basic screens. In particular,
screening in 'sensitized' backgrounds harboring weak
mutations in a pathway of interest have been useful for
identifying additional pathway genes whose loss of func-
tion phenotypes are weak in a wild-type background.
However, while such enhancer screens can identify genes
that increase the phenotypic penetrance of a specific
mutation, there is often little or nothing known about the
e
Classical Screens with Next-Generation
Sequencing
Genetics Leaps Forward
In a classical genetics screen in the worm, generating and
picking mutants is easy: identifying the causative mutation
(and hence the genotype
e
phenotype inference) is not
or
e
e
