TreeMap 3

Download TreeMap 3 here

Please send me an e-mail when you do so, so I can apprise you of developments and see how many people are using TreeMap!

Licensing

The code makes use of the free iText library, kindly provided by Bruno Lowagie and Paulo Soares. Their software, and hence TreeMap3, is distributed under the MPL.

Support

This work is supported in part by the ARC.

Summary

This is a draft document describing the new software TREEMAP3, which has been a very long time coming. I have been spurred on to re-write the thing in Java, partly because I like programming in new languages (to me) but mostly because I was greatly embarrassed to learn that some of my most respected colleagues have been holding on to old Macintosh computers that run Classic, just so they can run TREEMAP2.

I've just received a very kind request for TM3 and I thought I may as well just upload the development version for people to have a look at, and this may work as further incentive (coupled with their polite urgings!) to finish it. Thus, this is NOT a final version; it does NOT do all the things you want it to do, but it will, and it will do more. And at this point I am very happy to entertain suggestions on what other features and things could go in.

So far, TM3 has the following functionality, which was already in TM2:

reading in a TreeMap input file and displaying the tanglegram;
constructing a jungle, subject to constraints;
estimating the number of randomized parasite trees that map as well into the host tree as does the original parasite tree;
draw the maps - that took Rod Page and me a while to do nicely in the first version and it hasn't been a priority until recently;
print figures out: this uses the iText libraries as mentioned above.

It does not (but will!)

deal with timing information

New things

TM3 is written completely differently from TM1 and TM2. For a start it's in a different language, Java, while TM1 and TM2 were in C++. I like both these languages, but have opted for Java for its promise of portability, and the comparative ease of GUI development.

TM3 also uses the concept of the liana, which I introduced in my "Principles of cophylogenetic maps" book chapter, available from http://www.springerlink.com/content/61023248h5873375/.

TM3 will use lianas to handle reticulate phylogenies, and my aim is to get it to handle multi-host parasites this way too.

It now outputs a jungle constructed using the constraints given to the tanglegram (using the admittedly horrible interface as yet), in the "dot" format. You can visualise the graphs using dot or GraphVis http://www.graphviz.org/doc/info/lang.html.

New things I've already put in

I finally grew sick of the annoyance of having to untangle tanglegrams by hand, and wrote a (very silly) heuristic to do it for me. It's the "Untangle" button. If you hold down the "alt" (a.k.a. "option" on my keyboard) key it tangles it all up.

Things I've just broken

The menu of tree drawing styles at the top right of the tanglegram pane. Those icons were there before, I'm sure of it!

Very cool new things I hope to include

These will be good.

Likelihood, using new techniques in development
Preferential host switching [1]
(Bayesian) parameter estimation
Failure to diverge events

References

[1]: M. A. Charleston and D. L. Robertson. Preferential host switching by primate lentiviruses can account for phylogenetic similarity with the primate phylogeny. Syst Biol, 51(3):528-535, Jun 2002.