HyPhy message board - [ADDITION]Recombination: Single Breakpoint

	Welcome, Guest. Please Login

Home

Help

HyPhy message board › HYPHY Package › Feature Additions › [ADDITION]Recombination: Single Breakpoint

(Moderators: Sergei, Simon)

‹ Previous Topic | Next Topic ›

Pages: 1

Send Topic

[ADDITION]Recombination: Single Breakpoint (Read 1820 times)

Sergei

YaBB Administrator

Offline

Datamonkeys are forever...

Posts: 1658
UCSD
Gender: male

[ADDITION]Recombination: Single Breakpoint
Feb 2^nd, 2005 at 11:40am

As per Simon's request I have added a simple standard batch file (available with builds after Feb 1st) to look for a single recombination breakpoint.

The algorithm is simple:

Using NJ (with TN93) reconstructed trees we

Fit a single tree to the whole alignment, using a user selected model
Try to break the alignment at every variable site and fit two separately inferred trees (again, NJ) using the same model (with global parameters shared, but branch lengths inferred independently).
Measure goodness of fit by AIC

If the analysis finds a putative breakpoint, more rigorous tests (e.g. KH test - a standard analysis) can be used to assess the significance of phylogenetic incongruence.

This analysis can compensate for rate variation (using an appropriate substitution model) and use codon models for cDNA.

Associate Professor
Division of Infectious Diseases
Division of Biomedical Informatics
School of Medicine
University of California San Diego

WWW

IP Logged

Simon

Ex Member

Re: [ADDITION]Recombination: Single Breakpoint
Reply #1 - Feb 11^th, 2005 at 9:30am

Dear Sergei,

Seems to work well. A few suggestions:

1. Rather than fit a model as part of the analysis, could a previously fitted model be loaded instead? Perhaps the recombination analysis could also be added to the result processor module.
2. More output would be nice: how about a table with the breakpoints, the AIC, the delta AIC, with optionally, the trees for each partition, a KH test for the significance of the difference in trees, and some tree distance measures. It might also be good to tally the sites that are contributing to the conflicting phylogenetic signal, but I'd have to think a bit how best to present these.

Best,
Simon

IP Logged

Pages: 1

Send Topic

‹ Previous Topic | Next Topic ›

« Home

‹ Board

Top of this page