We have constructed a neighbor joining tree in MEGA with few eukaryotic protein sequences. According to BLAST result, our query shared 96% identity with Organism A and 93% identity with Organism B. But in the constructed phylogenetic tree, query comes closer with Organism B (93%) whereas organism A (96%) comes in a separate clade. I have tried with different set of Gap opening and Gap extension values. Still Organism B ( lesser identity) comes together with the query sequence.
What is the reason behind this? Could anyone explain?
What model did you use? Jukes-Cantor, or Tamura-Nei?
The phylogenetic tree is not based on strict overall similarity, so your results are not at all unreasonable. The tree is based on an actual model of evolutionary change across the sequences. Since not all nucliotide changes in a sequence are equally likely, the overall similarity does not tell you much at all about patterns of evolution. Two sequences can be highly similar, yet quite distantly related in phylogeny.
What model did you use? Jukes-Cantor, or Tamura-Nei?
The phylogenetic tree is not based on strict overall similarity, so your results are not at all unreasonable. The tree is based on an actual model of evolutionary change across the sequences. Since not all nucliotide changes in a sequence are equally likely, the overall similarity does not tell you much at all about patterns of evolution. Two sequences can be highly similar, yet quite distantly related in phylogeny.
agree with Micheal Black, evolution doesn't mean similarity. probably your sequences evolved more closely with organism B rather than A. also refine your alignment.
I´m also agree with Michael Black and Zareen Fatima. However I suggest you use either of these bioinformatics tools for a better presicion of your results.
http://www.ebi.ac.uk/Tools/msa/tcoffee/
https://www.ebi.ac.uk/Tools/msa/muscle/
I agree with all answers above. For alignment, I quite like MAFFT myself, but agree that t-coffee and MUSCLE are fast and probably better than BLAST as a step before phylogenetic analysis. Further than that, I wish to add that a bootstrap analysis coupled with a maximum parsimony or maximum likelihood phylogenetic tree will also provide you with branch support further adding to your confidence in the relationship. In any case it will depend a lot on what question you want to answer with the phylogenetic tree in the first place.
- Badges
- Science topic
- Similar topics
- Bioinformatics
- Bioinformatics and Computational Biology