AWESOME computer program (Populus) A couple of weeks ago, I picked up a truly great ecolog
AWESOME computer program (Populus)
A couple of weeks ago, I picked up a truly great ecology/evolution
simulation program. I thought some of the folks here might like to
hear about it.
The program is called Populus and is written by some folks at the
Ecology, Evolution and Behavior department at the U. of Minnesota.
Populus includes over 40 (and the number keeps rising) simulations
of ecology and evolution. Ecology simulations include both discrete
and continuous models of exponential and logistic growth.There is also
a model of age structured growth where you determine the probablity
of survival of each age class in a population and what the average
reproductive output is at each age (this is called the l(x)m(x)
schedule in ecology). The output includes the age structure of the
population at each interval, overall growth rate, reproductive value
of each age class and some other neat stuff.
There are also models of competition between two species as well as
numerous predator-prey models (Lotka-Volterra is there, of course, with
all the functional response options (options that let you determine how
capable predators are at capturing prey at different prey densities).
In addition, there is a "difference engine" that allows you to set up
your own simulations. Just type in your formulas and the program solves
the equations for you and gives you whatever output you want.
The best parts are (in my extremely biased opinion) the evolution
models, especially the models of drift and selection.
Drift can be modelled via a Monte Carlo simulation or using the
Markov probability method. There are several models of natural selection,
these include autosomal selection (you can model "normal" selection
here including selection against recessive lethals and balancing
selection (selection for the heterozygote), X-linked selection
(females have 2X chromosomes, males have one -- normal autosomal
models of selection amongst alleles don't work.), group selection
(you can see mathematically why it doesn't occur very often), sex-
ual selection, frequency-dependent selection and density-dependent
selection and a few others.
WARNING TANGENT ALERT -- I asked the author if he is planning
on adding models that incorporate "drive" or segregation distortion
alleles -- alleles that show up in more than 50% of all gametes.
Lots of cool things pop out the math, especially if you add in
antagonistic sex alleles (the allele has a different fitness and
drive capability in the different sexes). I just heard a talk on this
the other day and I derived (derove?) all the math later that night at
the pub -- it's really cool (not to mention simple) math. In any case,
some drive stuff might be added in the future. -- END TANGENT
There is also the (incredibly stupid, IMHO) "methinks its like a
weasel (its "woozle" in Populus)" simulation.
There's some neat quantitative selection stuff. You can start out with
a distribution of phenotypes, set the heritability and environmental
components of the trait and pick what percentage of the population
will be selected for breeding (for instance the top 10%) and you
can watch the phenotype distribution change each generation.
Two visually impressive models are: the Nicholson-Bailey model with
spatial structure and the 3 allele haploid host-parasite model.
NB is an unstable predator-prey model (the system eventually "crashes"
under most starting conditions).The NB with spatial structure model sets
up a grid, runs the NB simulation in each sector, then disperses
predators and prey to surrounding sectors (all these variables can
be changed). You can watch the prey/predator density of the cells in
the grid change with time (lots of cool osscilations, spirals, chaos
and etc.) Way cool.
The 3 allele haploid host-parasite model has three host types (alleles)
and three parasite types (predator alleles) -- this could be used to
model viral infection of bacterial populations for instance. The
first parasite "prefers" the first host type (you set the fitness
cost to landing on the wrong host), the second prefers the second
and so on. The model ouputs the change, in three dimensions (shading
on the sides of the cube), in frequency of the host and parasite
alleles. The host alleles are a single point (in three dimensions)
that circles around the space and is "chased" by the parasite
point. It looks extraordinarily cool.
I can't say enough great things about this program, it's awesome.
And, it's only 10 bucks. The author is giving it away at cost (and
encourages you to let interested parties copy it.) I wrote the
head author and asked if I could recommend it here and he said sure.
He wanted me to mention that many of the options may be boggling
to non-biologists. However, I could point out that each model has
several screens of background information explaining it and points
to references to the primary and secondary literature where more
info can be obtained. A basic ecology text (like Ricklefs) and
a population genetics text (like Hartl and Clarks) could clear up
almost anything presented in the program -- although it contains
a few things I hadn't even heard of until now.
The program runs on any IBM compatible machine with almost any graphic
display (Hercules,CGA, VGA, EGA and IBM8514). It only requires
500Kb of RAM and runs under DOS. No Mac version is available (or
planned). You can get a copy by sending 10 bucks (payable to U of
Dept. of Ecology, Evolution and Behavior
University of Minnesota
318 Church St. SE
Minneapolis, MN 55455-0302
Chris Colby --- email: firstname.lastname@example.org ---
"'My boy,' he said, 'you are descended from a long line of determined,
resourceful, microscopic tadpoles--champions every one.'"
--Kurt Vonnegut from "Galapagos"
E-Mail Fredric L. Rice / The Skeptic Tank