Macdonald Lab
Genetics & Genomics
Genetic basis of complex traits
Quantitative, or complex traits are controlled by many, potentially
interacting genetic and environmental factors. Many human diseases as
well as traits of ecological significance are likely to be complex.
However, despite
their importance, progress towards identifying the genetic loci
contributing to variation in complex traits has been slow. Research on
model organisms has proven to be a valuable approach with which to
dissect general biological phenomena. The goal of my research is to
understand the molecular genetic basis of complex traits using genomic
and analytical tools, together with the powerful resources available
for the
elite model genetic organism Drosophila
melanogaster.
This system holds
promise to answer fundamental questions about the evolutionary genetics
of complex traits: What types of genes contribute to standing variation
in complex traits? What are the frequencies and effects of
polymorphisms in these genes? Are causal sites generally coding, or are
they present in regulatory regions? Is there epistasis between causal
sites, either within candidate genes, or between genes in the same
biochemical pathway?
Association
Mapping
The modern paradigm for the genetic analysis of complex traits is
association mapping. This approach has the potential to uncover the
actual site(s), or QTN (Quantitative Trait Nucleotides) influencing
trait variation in nature. However, to ensure that subtle-effect QTN
can be
detected, very large sample sizes are required, and many SNPs (Single
Nucleotide Polymorphisms) must be genotyped. In the lab we have various instruments that allow us
to carry out large, powerful association studies.
Association studies must
generally be undertaken on a gene-by-gene basis, as the level of
polymorphism, and the lack of long-range linkage disequilibrium between
SNPs, generally preclude whole-genome association mapping. Fortunately,
Drosophila
researchers often have access to good candidate genes for
traits of interest. The Enhancer
of split
gene complex is a strong candidate
to harbor alleles with quantitative effects on bristle number
variation. We performed an experiment to assess the contribution of
genetic variation at Enhancer
of split to bristle number variation in a
sample of 2000 wild-caught D.
melanogaster (Macdonald
et al. 2005).
Laboratory vs.
Nature
In model organisms, genetic factors underlying phenotypic variation are
commonly identified under laboratory conditions using genetically
manipulated strains. Are these factors responsible for conferring
phenotypic variation in nature? Conclusions about the ecological and
evolutionary relevance of laboratory-identified genetic factors assume
that effects measured in the laboratory are similar to those seen in
nature. Some of our work has shown that strong laboratory-identified
associations may not replicate in nature (Macdonald
& Long 2004).
The generality of this result is unclear and awaits further
work.
However, it does suggest that it may not be straightforward to extend
the results of laboratory studies to natural populations.
In silico
Identification of Functional Sites
Since only a tiny percentage of the segregating polymorphism within a
species is likely to be functional and contribute to phenotypic
variation, is it possible to enrich the subset of genotyped SNPs for
those most
likely to be functional? I have performed surveys to examine how
primary sequence data might be used to predict likely functional SNPs.
It is of particular interest to functionally
annotate noncoding DNA, as much phenotypic variation is
thought to
be controlled by cis-regulatory
changes. I sequenced noncoding DNA from regions across the D. melanogaster
genome, and used a variety of analytical tools to "tag" regions as
nonneutrally evolving: Compared to SNPs in regions showing no evidence
of past selection, SNPs in tagged regions are stronger candidates to be
functional (Macdonald
& Long 2005).