Hello, I am working with a two genes sequences of octopus (CO1 and NADH1) and want to generate phylogenetic tree from it. I was looking for some suggestion about how to construct the best tree from it and got it from here
https://www.fifthdimension.jp/documents/molphytextbook/molphytextbook.ja.html
problem is, it is in Japanese and google translate did not help much. Can someone help me find the similar manual for constructing a tree using kakusan? or maybe a better way to do it?
Thank you
As I know Kakusan is a partition finder and model selection software (sorry if I'm mistaken). As a beginning you can use MEGA (https://www.megasoftware.net/) easy to use and has built in model selection procedure (alternatively you can use jmodeltest: https://github.com/ddarriba/jmodeltest2) and has several phylogeny inference module like ML, NJ and ML.
Kakusan can use different models for different genes or "partitions", but both COI and NADH1 are highly conserved protein coding mitochondrial genes, so they should have very similar evolution within one group of organisms such as cephalopods. If you had one structural RNA gene and one protein-dding gene, the use of different models might make more sense. But you can test each data set with ModelTest or PartitionFinder to double check on this.
More often, for protein-coding genes, it makes more sense to treat first, second and third codon positions differently (partition by codon site) than to treat different genes differently (partition by gene).
http://www.robertlanfear.com/partitionfinder/
https://www.hiv.lanl.gov/content/sequence/findmodel/findmodel.html
DAMBE, MEGA, and dozens of other phylogeny packages can be used to construct trees by NJ, ML and other methods with as many different models as you like. Or you can use web services such as IQ-tree (Maximum Likelihood with any model you like).
http://dambe.bio.uottawa.ca/DAMBE/dambe.aspx
https://www.megasoftware.net/
http://www.iqtree.org/
All of this is probably a bit of overkill for 2 mitochondrial protein-coding genes. In my experience it is far better to put effort into good data sets than to beat a limited data set to death with more computer power.
Thank you Prof. Csaba and Prof. Brian,
It is really helpful. I will try to do it by MEGA and learn partition by codon site and gene. Thank you so much
- Badges
- Science topic
- Similar topics
- Psychology
- Mental Processes
- Learning