Time Scale For Bayesian Skyline Plot

[FrankBurbrink]

First off, I wanted to say that this is one of the most useful
phylogenetic programs ever developed in terms of research and teaching
students. Thanks.

I have a question concerning the time scale (x-axis) used for Bayesian
Sky Line (BSL) plots feature.

If I 1) enter in a mtDNA sequence dataset for a population, 2) set the
root height of the population for some value (1.5 million years) as a
lognormal prior as dated from another study, and 3) choose Bayesian
Sky Line plots for demographic estimation, then how is that axis
scaled against population size (Ne) on the BSL? That is, for a
population demographic analysis would I have to alter the x-axis in
time provided by Tracer to account for generation time or is that
scale absolute and should be used?

Basically, I want to determine if population growths and bottlenecks
follow certain glacial cycles in North America and I must be certain
that I am calculating that X-Axis correctly.

Thanks a million!!!

Frank

[FrankBurbrink]

[archdev]

Dear Frank,

Units will be the one you'd used for calibration. So if you had
entered 1.5 to stand for 1.5 million years, units will be in millions
of years. If you had entered all the digits (1500000 or 1.5E6 years),
then units will be in years.

Hope that answers your question.

archdev

[Simon Ho]

Hi Frank,

In addition to archdev's answer, the y-axis in your analysis measures
(Ne.t), the product of the effective population size (Ne) and
generation time in years (assuming that your time scale is given in
years).

Simon

[Andrew Rambaut]

Frank,

On 18 Jul 2007, at 23:21, FrankBurbrink wrote:

> First off, I wanted to say that this is one of the most useful
> phylogenetic programs ever developed in terms of research and teaching
> students. Thanks.

Thanks - I never use anything else!

> I have a question concerning the time scale (x-axis) used for Bayesian
> Sky Line (BSL) plots feature.
>
> If I 1) enter in a mtDNA sequence dataset for a population, 2) set the
> root height of the population for some value (1.5 million years) as a
> lognormal prior as dated from another study, and 3) choose Bayesian
> Sky Line plots for demographic estimation, then how is that axis
> scaled against population size (Ne) on the BSL? That is, for a
> population demographic analysis would I have to alter the x-axis in
> time provided by Tracer to account for generation time or is that
> scale absolute and should be used?

The time scale (x-axis) of BSP & the other demographic plots will be in
the same units you used to calibrate the tree (in your case, million
years).
The population size (y-axis) will be a product of Ne and generation time
(in your calibration time units).

Andrew

___________________________________________________________________
Andrew Rambaut
Institute of Evolutionary Biology University of Edinburgh
Ashworth Laboratories Edinburgh EH9 3JT
EMAIL - a.ra...@ed.ac.uk TEL - +44 131 6508624

[FrankBurbrink]

Thanks Guys!

I thought that was the answer, but I wanted to be sure before I submit
my current paper on population expansion in racer (a snake) lineages!

Frank

> EMAIL - a.ramb...@ed.ac.uk TEL - +44 131 6508624