Title: The Role of Epistasis in the Control of Complex Traits
1The Role of Epistasis in the Control of Complex
Traits
- Örjan Carlborg and Chris Haley
- Roslin Institute
2Epistasis
- Epistasis interactions between alleles at
different loci - Relatively neglected in quantitative genetics
- More attention in qualitative genetics where
several types of epistasis has been found for
e.g. coat colour phenotypes
3Examples of classic epistasis
Hybrid dysgenesis
4Epistasis and QTL
- Most QTL mapping studies focus on detection of
marginal effects assuming absence of epistasis - Assumption of no epistasis based on low estimates
of epistatic variances from quantitative genetic
studies - Some studies have searched for epistasis between
QTL with significant marginal effects
5Why consider epistasis?
- Similar cellular mechanisms affect quantitative
and qualitative traits - Increased power to detect QTL which mediate their
effects through interactions - Epistasis could bias estimates of additive and
dominance effects - Epistasis could help in the interpretations of
QTL mapping results
6Methodology to detect epistasis
- Need procedure to detect QTL with marginal
effects as well as entirely epistatic QTL - One option (Carlborg and Andersson, 2002)
7Increased power to detect QTL
- Possible to detect novel QTL pairs with very
small or no marginal effects at all possible for
the individual QTL - Increased support for QTL with close to
significant marginal effects
8New epistatic QTL detected
- Some QTL can not be detected by their marginal
effects - E.g. QTL pair affecting hatch weight in chicken
9New epistatic QTL detected
- Simultaneous mapping of epistatic QTL increases
significance for some QTL with minor marginal
effects
10Bias in estimation of QTL effects
- Estimates of marginal genetic effects are the
average effect of the alleles at a locus across
all alleles at all other genomic loci - By identifying interactions between loci one can
obtain additional information about e.g. - In which genetic background a QTL has its most
influential effects - How further studies should be designed to
maximise the chance to replicate a QTL
11Bias in estimation of QTL effects
12Additional variance explained
- The estimates of the relative amount of variance
attributable to epistasis depend on the genetic
model used in the study - The simplest and most commonly used model is that
of Mather and Jinks - Alternatives include the Cockerham model
13Additional variance explained
- Epistasis explains a considerable portion of the
genetic variance (Mather-Jinks model)
14Genetic mechanisms of epistasis
- By understanding more about the nature of QTL
interactions, one can - Be more precise when selecting candidate genes
- Better select the strategy for replicating a QTL
- Get higher efficiency in marker assisted
selection (MAS) - Need to explore Genotype-Phenotype patterns for
epistatic QTL
15Genetic mechanisms of epistasis
- Chicken F2 intercross between growth and layer
selection lines - 21 epistatic QTL pairs detected
- 17 pairs could be associated with one of 4 types
of epistasis
16Genetic mechanisms of epistasis
- Chicken F2 intercross 4 common types of
epistasis
17Conclusions
- Epistasis is abundant
- Epistasis increases power in QTL mapping
- Novel QTL detected and more variance explained
- Epistasis might be needed to replicate QTL
findings and clone QTL - Can not understand control of complex traits
without considering epistasis
18Acknowledgements
- Roslin Institute
- Dave Burt, Paul Hocking, DJ deKoning
- Swedish University of Agricultural Sciences
- Leif Andersson, Per Jensen, Susanne Kerje, Karin
Schutz - The Knut and Alice Wallenberg foundation
- NGSSC
- BBSRC