Hi,
I use datamonkey for recombination analysis. GARD indicated no evidence of recombination, but SBR(Single Breakpoint Recombination) methods indicated that there was a breakpoint at 193, with a support of >99%, whether recombination inferred using AIC or cAIC.
I am not sure about the discordance. Is there any recombiantion in my sequence?
THanks in advance.
kevintz

Thanks for your reply. I still cannot figure it out.
The following is the result page.
Multimedia File Viewing and Clickable Links are available for Registered Members only!!  You need to Login

Another question, the selected best fit model by jmodeltest is TIM2+G(010232) for my dateset (20 sequences with 1485bp long), but by the datamonkey server the best model is TrN model (010020), How to treat with the different results. And when I try to use 010020 model by changing one AC to CT according to the model selection result page, and run SBP or GARD, the rusults showed the model is 010040. I wonder how to fix it.
THanks a lot

Hi Kevin,

You are referring to the alignment from section 3.10 in Multimedia File Viewing and Clickable Links are available for Registered Members only!!  You need to Login, namely Multimedia File Viewing and Clickable Links are available for Registered Members only!!  You need to Login right?

For this alignment, the key is to include rate variation in the analysis set up (see the attached screen shot). Without rate variation, the boundaries between more/less variable stretches of the gene can be identified as recombination breakpoints. If you use rate variation, the results are mostly congruent with the book chapter (Multimedia File Viewing and Clickable Links are available for Registered Members only!!  You need to Login). We've revised the algorithm and HyPhy optimization routines a bit since the book was published, and since the procedure is also stochastic, some 'wobble' in the location of inferred breakpoints is expected (but look at the confidence intervals as well). Notice that the new GARD output includes breakpoint-by-breakpoint phylogenetic incongruence evaluation.

As for the influenza example, the fact that GARD infers breakpoints, could point to heterotachy or other process which generate discordant phylogenetic signal but are not recombination events. I think all the book says is that intra-gene recombination is thought to be rare in Influenza, but

Sorry for the confusion.

Sergei