Phylogenetic Analyses Of Sequences Containing Mixed Bases
Dear all, I have a dataset of DNA sequences containing a large amount of mixed bases. Most phylogenetic programs I have used so far do not take into account positions where at least one sequence has a mixed-base (edit : or I failed to use the right setting for this utilization). This results in a heavy loss of information in my case.
Therefore, I am now searching for phylogenetic programs or tips to construct NJ, ML and Bayesian trees while considering mixed bases as informative characters.
Do you have any advices ?
hugo
1 answer
Hugo,
Lots of software will handle ambiguous bases (RAxML, MrBayes,dnadist...) , but there are some things to think about.
- Are you sure you are looking and mtDNA heteroplasmy, and not a nuclear pseudo-gene (NUMT) that is being co-amplified
- If you really do have heteroplasmy you don't have ambiguity, you have two mitochondrial haplotypes. Munging that signal into a single consensus can (but won't necessarily) have unwanted consequences (see White et al, 2008 for some background on this).
Log in to answer this question.
More posts like this
what do you mean by a mixed base? Do you mean IUPAC code (e.g. Y for pyrimidine?)?
My samples exhibit mtDNA heteroplasmy. This results in some individuals producing mixed signals with sanger sequencing at a few positions. I do use IUPAC code for these positions.
Have you tried MEGA5?
I did. MEGA can handle sequences with ambiguous bases. However, it seems that it deletes positions with ambiguity for NJ and ML tree constructions. Maybe I do not set it properly ?
This is my opinion, but I think MEGA is a horrible program. I am not aware of any phylogenetic analysis programs that DON'T accept ambiguous bases. Information in user manuals will explain how each respective program codes for ambiguous bases or unresolved sequencing calls.
Ok. Thanks. I will study this further.