Transcript Stephan Hoyer.

Stephan Hoyer
Physics 120 seminar
March 28, 2008
EVOLVABILITY
What is evolvability?
 Two definitions:
 A biological system is evolvable if its properties
show heritable genetic variation, and if natural
selection can thus change these properties
 A biological system is evolvable if its properties
can acquire novel functions through genetic
change, functions that help the organism survive
and reproduce
“Evolvability” at different scales
Massimo Pigliucci. Is evolvability evolvable? Nature Reviews Genetics 9, 75-82 (January 2008)
Extra eyes on Drosophila
Modular
Robust?
Halder, G., P. Callerts, and W. J. Gehring. Induction of ectopic eyes by targeted expression of the
eyeless gene in Drosophila. Science 267, 1788-1792 (1995).
Robustness
 A biological system is robust if it continues
to function in the face of perturbations
 For living things:
 Genetic robustness
 Environmental robustness
 Different levels:
 Changes in genotype not reflected in phenotype
 Overall fitness… to survive, reproduce, etc
Does robustness help evolvability?
 Robustness hurts:
 Less phenotypic variation => Less selective
pressure => less evolvability.
 Robustness helps:
 Versatility: for other non-robust traits
 Neutrality: robustness means more silent
mutations to set stage for important ones.
 As Capacitors: when robustness breaks, hidden
variation is released
Evolvability in fitness landscapes
Q: What’s wrong with this picture?
A: Far too few dimensions!
Neutral spaces
 Visualizing fitness landscapes gives you way
too few dimensions!
 Peak shift problem: how do you move
between peaks?
 High dimensional (100-1000D) landscapes
are qualitatively different
 Neutral spaces allow for lots of variation
 Techniques from physics: percolation theory and
simulation
Gavirlets, S. A dynamical theory of speciation on holey adaptive landscapes. Am.
Nat. 154, 1-22 (1999).
Why robustness?
 Clearly life actually is very robust
 Easier to get to by chance most stuff exists
in the large neutral space
 Natural selection can select for robustness
in neutral spaces
Modularity
 Modularity helps evolvability?
 Are modular systems robust?
 Altering one module is less likely to disrupt
others
 But that one module can still disrupt everything
if it’s essential
 Also requires redundancy, most not in genes, but
in alternate methods
How does evolvability evolve?
 Clearly there are big steps in evolvability,
e.g.




Multi-cellularity
Animals
Appendages
etc
 Is evolvability a side-result or the end-goal
of evolution?
 Brings up issue of evolution at different levels:
for individuals or for species?