Dear colleagues,
We have a new 3-year PDRA position open for a project using statistical
mechanics and algorithmic information theory to investigate evolutionary
patterns in biological development.
The deadline is midday (UK time) Wed 5 May, and the application link
is here
I would appreciate it if you can pass this on to anyone you think might
be suitable for this post.
Further details on the project can be found at:
http://www-thphys.physics.ox.ac.uk/people/ArdLouis/evolution.shtml
below:
many thanks,
Ard Louis
<<>> LONGER DESCRIPTION BELOW <<>>
*PDRA on Arrival of the fittest in development: *
« Nothing in Biology Makes Sense Except in the Light of Evolution. » wrote
the great naturalist Theodosius Dobzhansky
to really understand evolution, a stochastic optimization process in an
extremely high dimensional space[1], will require techniques from
statistical physics.
Darwinian evolution proceeds by two steps. First, random mutations
generate new heritable phenotypic variation. Second, the process of
natural selection ensures that phenotypes with higher fitness are more
likely to dominate in a population over time. Research in evolutionary
theory has mainly focussed on the second step, natural selection. Much
less is known about the first step, the arrival of variation. For a
popular introduction to this topic, see the charming book: Arrival of
the Fittest
Wagner.
Much of our work in this area uses theoretical tools of statistical
mechanics and algorithmic information theory, together with computer
simulations, to study genotype-phenotype maps, which give access to the
structured role of variation in evolution. As an example, see this
paper on phenotypic bias [2], where we quantitatively predict the
frequencies with which RNA structures appear in nature, without taking
natural selection into account. Instead, a very strong bias in the
arrival of variation dominates over selective pressures, a non-ergodic
effect we call the arrival of the frequent [3].
We have recently shown that such strong bias in the arrival of variation
also explains structural patterns in protein quaternary structures and
gene regulatory networks [4]. The big open question for this postdoc is
whether such patterns can be observed beyond these molecular phenotypes,
at the larger scales of the evolution of development. Arguments based
on the coding theory from algorithmic information theory [5] suggest
that strong bias may hold also for some aspects of development, but that
first needs to be established (or falsified) explicitly.
In this project, you will use a wide range of techniques to explore the
mapping from genotypes to phenotypes in models of development, and to
study the adaptive evolutionary dynamics on these landscapes. There
are close analogies to the question of why overparameterized deep neural
networks generalize so well [6], and part of the project may include an
exploration of these commonalities.
This is a challenging interdisciplinary project. Strong quantitative
skills and a proven track record of creative and successful independent
research are the most important qualities we are looking for.
Experience with biological physics is a plus, but not a requirement.
Finally, for an overview of our work in this area see this recent talk:
[1] Contingency, convergence and hyper-astronomical numbers in
biological evolution
A. A. Louis, Studies in History and Philosophy of Biological and
Biomedical Sciences *58*, 107 (2016)
[2] Phenotype bias determines how RNA structures occupy the morphospace
of all possible shapes
Dingle, F. Ghaddar, P. Sulc, and A. A. Louis, bioarxiv
[3] The arrival of the frequent: how bias in genotype-phenotype maps can
steer populations to local optima
S. Schaper and A. A. Louis, PLoS ONE 9(2): e86635 (2014)
[4] Symmetry and simplicity spontaneously emerge from the algorithmic
nature of evolution, I. G. Johnston, K. Dingle, S. F. Greenbury, C. Q.
Camargo, J. P.K. Doye, S. E. Ahnert, and A. A. Louis
[5] Input–output maps are strongly biased towards simple outputs
K. Dingle, C. Q. Camargo and A. A. Louis, Nature Comm. *9*, 761 (2018)
[6] Deep learning generalizes because the parameter-function map is
biased towards simple functions
Valle Pérez, C. Q. Camargo, A. A. Louis ICLR (2019)
—————————————————————————————————–
Prof. A. A. Louis Rudolf Peierls Centre for
Theoretical Physics
ard.louis@physics.ox.ac.uk
University
phone: +44 (0)1865 273994 Clarendon Laboratory, Parks Rd,
fax: +44 (0)1865 273947 Oxford, OX1 3PU, UK
http://www-thphys.physics.ox.ac.uk/user/ArdLouis/
—————————————————————————————————–