Hi everybody,
If a phylogenetic tree is reconstructed by maximum likelihood (ML) method, how the DNA sequences for one gene or genetic marker must overlap after alignment step ? There is a treshold to have a good phylogenetic reconstruction to respect ? If yes, can you indicate me with reference to read and learn.
Thanks you and best wishes
I don't think there's an absolute limit. Given that the phylogeny is really a tree, topological support depends on two things: sequence length and sequence variability. Hence, a very short sequence can be placed with accuracy within a phylogeny if it is (near) identical to exactly one other, longer sequence and they contain enough synapomorphies.
The trick to maximise the phylogenetic resolution of a data set with lots of missing data is to run a Bayesian dating analysis. Such analyses have a prior on branch length (i.e. tips all align to time=0), which considerably limits the number of places where branches that are underpinned by few characters can attach to. For such data sets (e.g. ITS) I use BEAST as a defaut, which generally gives a far better outcome than e.g. MrBayes, ML or MP.
Thank you for your answer, so I need to make a first analyse with BEAST, I do't really understand why such analyses have a prior on branch length ?
And what is a ITS data set ?
Dating analyses always have a branch length prior, because the tree is ultrametric (all the tips were collected 0 Ma ago), and a molecular clock prior. A sequence with many autapomorphies is expected to coalesce more deeply in the tree than a sequence with few autapomorphies.
ITS1 and ITS2 are the ribosomal internal transcribed spacers 1 and 2.
- Badges
- Science topic
- Similar topics
- Biological Science
- Evolutionary Biology
- Phylogenetics