Hudson Animator
 
The Hudson animator consists of a number of tools for visualization of the coalescent process under different assumptions.
   The principle of the animator is illustrated below for the case of the Coalescent with Recombination. Other case: Coalescent with Migration, Coalescent with Growth and Coalescent with Selection are explained in the last part of this document.

Coalescent with recombination
The principle is based on Hudson's (1983) algorithm which included recombination into Kingman's (1982) Coalescent process. The coalescent with recombination for a sample of genes is defined by exponentially distributed waiting times for either coalescent or recombination. Coalescent events merge two sequences into one, whereas recombination breaks up the sequence in two (corresponding to the two different sequences that recombined to form the sequence in question). Since recombination splits up sequences, the process results in a graph rather than a tree. This graph has been termed the ancestral recombination graph (Griffiths and Marjoram 1996).
   The Hudson animator illustrates how this ancestral recombination graph is constructed, including which time coalescent/recombination events happen, which sequences merge, where recombination breaks up the sequences, and when the most recent common ancestor (MRCA) of the sample is found for each part of the sequence.

From the ancestral recombination graph it is possible to extract the real coalescent tree for each point the the sequence. This is illustrated in the Trees part of the animator.

 
The Hudson Animator
The following is a screen dump of the animator:

 
View
Use the fans to swap between the ancestral recombination graph the trees extracted from the graph.

Controls
Use the control panel to control the animation. The controls are the following:


Rewind

Step Back

Pause

Play

Step

Go to end

The animator operates similar to a video recorder. Once an animation is loaded into the animator, it is possible to 'Play' it, 'Pause' it, 'Rewind' it and so on. Note that the controls do not change the ancestral recombination graph, only the visualization of it.

Status
The status bar shows useful information about the current state of the animator and gives help on some of the components in the animator (eg. when moving the mouse pointer across certain components).

Legend
Key to the different kinds of nodes in the animator.

Model Parameters
This panel allows you to change model specific parameters. The parameter panel for the coalescent with recombination model is the following:

  The model has two parameters:

n : Sample size
rho : Rate of recombination

Info
Information about nodes in the ancestral recombination graph. This includes the time of nodes (measured backwards from time 0 in 2N generations, N being the effective population size) and other kinds of model specific information. The information is changed whenever the mouse pointer moves across a node in the graph.

In the recombination model only two types of nodes exist: Coalencent nodes and Recombination nodes. The two types display a different kind of information, as illustrated in the following figures:

 
Coalescent info  Recombination info

Coalenscent nodes display information about the part of the sequence finding most rescent common ancestor (green), the part of the sequence still missing most recent common ancestor (red) and the part of the sequence that is trapped (white). Recombination nodes display information about the point on the sequence where the recombination takes place and the parts that remain after splitting the sequecne according to the recombination point.

To start an animation, first locate the Model Parameters panel and change the model parameters to reflect your desire. Next press the 'Recalc' button to generate a new animation tree based on the selected set of parameters. If all goes well the animation automatically starts running in the main animation window. Otherwise the Status panel will inform about the error.

 
Other models
Exponential growth
In this model the population size increases exponentially. The parameter panel is the following:
  The model has two parameters:

n : Sample size
exp : Rate of growth

Migration
In this model the population divided into two populations. Coalescent is only possible between sequences in the same population but it is possible for a sequence to "jump" to the other population (migration). The parameter panel is the following:
  The model has three parameters:

n : Sample size
M1 : Rate of jumps from population 1 to population 2
M2 : Rate of jumps from population 2 to population 1

Selection
This model is somewhat different from the other models. With selection a selection graph is first constructed. Then a tree is extracted from the graph by placing mutations on arbritrary branches in the graph and applying a set of reduction rules (the extraction process is not yet visualized by the tool). The parameter panel is the following:
  The model has three parameters:

n : Sample size
S : Rate of selection
U : Rate of mutation (average number)