I suspect your alignment is too short (see example in Science 310, 1910-1911 [2005]) but it could also be that your data have too many sites that are incompatible in a phylogenetic sense (i.e., different sites supporting conflicting trees). Try to analyze the data using IQ-TREE (its maximum likelihood program); I've found it very useful.

 Quality of the sequences perhaps, were they checked for chimeras for example? as this is a functional gene maybe you can try a phylogenetic analysis in the protein space? maybe the sequences you included for comparing yours were not appropriate? maybe you can try a maximum likelihood or bayesian method instead of NJ - which is based on distances only..  there are a lot of options here. I would go first by trying an analysis with associated protein sequences then look at the alignment and check visually if the alignment make senses...   when dealing with nucleotide sequences you should always take care of the forward/reverse complement issue.. maybe not all of your sequences are in the same direction... and so on... 

I hope that helps

---sram

I suspect that your sequences have too little signal. That is I think that either the sequences are too similar to one another to detect any tree-like pattern or the sequences are too noisy and too divergent. If you'd like to send the sequences, I could take a look and tell you what your problem is.

I suspect your alignment is too short (see example in Science 310, 1910-1911 [2005]) but it could also be that your data have too many sites that are incompatible in a phylogenetic sense (i.e., different sites supporting conflicting trees). Try to analyze the data using IQ-TREE (its maximum likelihood program); I've found it very useful.

I did an alignment of the trimmed sequences using Muscle and a  reanalysis using PHYML (as implemented in Seaview, a free and very good analysis package) and maximum likelihood methods - my usual method. The model is GTR+I+G.  

I attach the tree file and which shows a lot of very good aLRT support values and also shows which OTUs are virtually identical with respect to the sequence data.

Here I use shortened names simply to do a quick run to determine how much phylogenetic signal is contained in the sequences. Take a look and see what you think and send me questions.You can see from the data and the tree that good branch support defines a lot of relationships and also that a lot of specimens are nearly identical to one another and so have extremely short branches. 

So the sequence data looks pretty good and phylogenetically informative.

Patrice,

Seaview is indeed very good, but you should try to use the model-selection method (ModelFinder) implemented in IQ-TREE. You might find that GTR+I+G is not be best model of sequence evolution.

All the best,

Lars

Lars is quite right. PAUP's decision theoretic (DT, i.e..BIC+) methods choose the GTR+G model.

Let me know if you run into problems.x                          

I have been working with plants phylogenies. I used ever several markers (nrDNA and cpDNA). To align the sequences, I used Geneious for that with Muscle. Before you try to do the analysis, maybe try to do jModelTest to find the best model. Normaly I use GTR+G, but sometimes the best fit model can be other. Check the AIC for the 3 firsts best models. Did you analyse correctly the gaps from your alignment? In my phylogenies I use a lot of samples, with a outgroup and I do the three analysis, after check the results of jModelTest: PAUP, RaxML and MrBayes. You have to describe this in your methods data analysis. Resolve better the align and check eventually the errors and that could solve the problem of your bootstrap values. Good luck!

The mcrA gene as an alternative to 16S rRNA, this region has high variblility rate, such as insecions and deletions, so the nucleotide aliment is dificult. one option for this region is to make aminoacid aligment, first translate your nucleotides sequences to aminoacid with The Bacterial, Archaeal and Plant Plastid Code (transl_table=11), that is used for this region. next, do aligment with amino acid, the results for this aligment should be the protein with not stop ( * ) codons, if * appear in the traslation you should move to the second o third position in your nucleotide sequence, to find the open reading flame of your sequences. After that you can make a model test from your data, and use those results to filogeny reconstrucction, to improve the reability of your data. you can use MEGA software 

Good luck