I just ran phylogenetic tree analysis of 16S rDNA sequences (around 1,500 bp) of Bacillus spp. using MEGA-6 and used Maximum Likelihood method with Bootstrap 1,000 replications and got the tree as attached. I was wondering if I did anything wrong since the scale bar appear to be 5 rather than less than 1 as I usually have seen. How can I interpret this tree with a scale bar of 5?
All 16S rDNA sequences are in the same 5' to 3' direction.
However I have tried Neighbour-joining method, it gave a different result of phylogenetic tree with a scale of 0.5, but the values of confirmation level at many nodes are equal to 0. Is that possible?
Please advise me. Thank you in advance.
That does seem odd but the branch lengths also are integers. Is it because you have bootstrap values on the branches from the 1000 replicates? The values look more like parsimony branch lengths. Can you send the Mega output?
Hi
Please try these conditions on your phylogenetic tree analysis
Tamura-Nei model, uniform rates, complete deletion, select all codon positions, nearest-neighbour-interchange,
I hope it will help you.
Hello Vijitra,
It could be that those branch lengths are scaled by raw stepwise mutations or substitutions and not normalized by the length of the sequence.
Best Regards
you should check all sequences direction. Are all the sequence 5` to 3` direction? if to sequence were the same but one 5` to 3` and the other 3` to 5` the dendrogram in mega6 will be dfferent and scale will be high
All 16S rDNA sequences are in the same 5' to 3' direction. That should not be a problem. However I have tried Neighbour-joining method, it gave a different result of phylogenetic tree with a scale of 0.5, but the values of confirmation level at many nodes are equal to 0. Is that possible?
some of sequences that deposited in NCBI is not correct. try to select the reference sequences from blast result and see the author name
HI,
The scale bar in phygenetic tree represent two scales ie. Time and distance.
scale also depends upon degree of similarity between sequences is in MSA. They are far away from each other the value of scale increases.
Hi Vijitra,
I still find this puzzling after making trees myself from your data. But I can make some remarks to your question.
Disregarding the strange scale bar, the branch labels on the ML tree you have are almost certainly bootstrap values (percentages of the 1000 replicates). They cannot be branch length labels becuase they are not proportional to the length of each line. Note that some very short branches have high value numbers.
I attach my own ML tree made with Phyml (I did not maximize the model, jus used a quick HKY). This looks pretty similar to yours -except I have it rooted differently. I have support values as percentages and branch lengths as decimals.
With any tree you certainly can get zero branch lengths and zero support on a branch. In particular with your data, you might note that many, many of the sites are invariant because many of the tremendous similarity among some of the sequences. Similar sequences will group together but if they vary from each other at no sites or at very few, then the connecting branches will be zero or nearly zero.
If you were to send me the raw tree file (nexus or newick), I could look at it and decide for sure what the labels are.
I suppose you aligned all sequences in MAFFT before uploading into Mega?
Thank you so much everyone for your responses. The problem was resolved by alignment of all sequences using ClustalW which was not done previously before constructing a phylogenetic tree.
Oh good! That makes sense because for my reanalysis of your sequences I did not get a scale of 5 and I had aligned it with Muscle.
Hello Vijitra,
Did you sequence these samples yourself or did you download them from GeneBank? If you did sequence the samples make sure you edit your sequences and trim the ends where necessary. Also try to make sure that the sequences cover the same region and do not have a bunch of missing data (i.e. if some are partial coding and other are complete coding).
Best Regards.
- Badges
- Science topic
- Similar topics
- Bioinformatics
- Bioinformatics and Computational Biology