I have quite not trivial situation. I have a sequence which only slightly overlaps with the rest of alignment (all but 1 sequence in dataset have only few not very informative bp overlap, one sequence is complete mt genome therefore it overlaps with all sequences in alignment), but want to know its' position on a tree. Unfortunately, there are no chances to resequence it. Are there any tricks to do it?
How many sequences do you have in total and how long the region of overlapping?
Find the position of the sequence in mt genome is possible by alignment (including also RC) and by BLAST in case if your sequence larger than 20 bp.
I need to find position of one sequence (extinct population, old DNA) on a tree from recent taxa, overlapping with it in 200 bp. I aligned them together against complete mitogenome, which is available.
You can do it easily by putting your sequence into nucleotide BLAST and match suggested organism mt genome below. After this you obtain its location into genome.
If you already find overlapping regions you can try to recontruct any phylogeny based only on overlapping region. In such way you able to analyse only 200 bp by various phylogenetic software to reconstruct phylogeny. But I'm quite afraid that will be difficult to reconstruct phylogeny on so small fragment, but all depends in loci and variability
Thanks, did exactly what you suggested, but in this 200 bp region phylogenetic signal is too weak to build resolved tree. I dont have much experience with supermatrix approach, but the idea seems similar to it - I have perfectly overlapping sequences A, B and C, but (ABC) and (D) are overlapping only partly. Finally, there is E sequences overlapping with all of them, will it help to infer phylogenetic relationships? Should it be ML, Bayesian or Parsimony?
You may try to increase phylogenetic signal by excluding sequences not good overlapping. Also try to find outgroup to all your species and add to analysis. Than you can use different algorithms for constructing trees and compare results. Firstly try to use simple distant metods to preliminary estimation of your tree (NJ, UPGMA). These algorithms are fast and give you some general view. After this you can use ML, MP or Bayesian analysis or its combination. If you able to provide screen of your data it will be helpful to advise you something more exact.
Thank you Andrii, will try what you suggested. I think I will also read carefully about supermatrix approach, my case may be similar.
Found reference! "adding incomplete taxa can be viewed as adding information, rather than adding problematic missing cells to the data matrix" - Alan de Queiroz and John Gatesy, The supermatrix approach to systematics. Trends in Ecology and Evolution, 2006, 22, 1. 34-41. Bayesian inference must work fine in such case.
- Badges
- Science topic
- Similar topics
- Methodology
- Research Methodology
- Research Papers