I have a maximum likelihood tree (18S and COI) for a nematode family which contains 4 different genera. The bootstrap values were very low and all the 4 genera are intermingled in both trees with no separation per region or genus. Does this mean that the genera separation is not supported?
Hey Lidia! Low bootstrap values indicate that there is conflicting signal or little signal in the data set. This may be a problem in the alignment, as Chris suggested. In this case it could be due to an erroneous alignment, which often occurs when the sequences aligned are ambiguous or too diverse. Sometimes, different clades had different rates of evolution (more or less mutation when compared to each other). Also, the absence of intermediate taxa could be a problem. Alternatively, information may be simply absent or too low in your alignment. That is the case when a very conserved DNA fragment has been used and sequences are too similar to each other. Normally you cannot conclude much else from low bootstrap values.
Normally, if a node in the dendrogram is not supported (usually bootstrap below 50) you cannot trust the bifurcations and that is why usually these nodes are collapsed in a tree graph to form a polytomy. If a node is poorly supported (50 to anything around 85 - a bit a case of personal interpretation, I think) it is to be treated with care but there is some support at least. Anything above that value can be interpreted with confidence.
I guess that your COI tree will have problems predominantly. Here, due to the protein-coding nature of this gene, the alignment should not be a problem. However, the third codon position may lead to "background noise" in your data because of the saturated code and silent mutations. A way to improve your results may be to infer relationships again with the third codon positions excluded.
As for the robustness of the concepts of the genera you study: you may doubt it is, as you wrote, they rely on poor descriptions. Think about the apomorphic morphological characters: are they complex characters, or is there evidence that they may have evolved multiple times independently? Usually, a taxon defined by more and more complex characters is more robust. I would not accept my molecular results blindly, but investigate, if they are plausible given alternative data (morphology). If so, robustly supported clades in your dendrogram that show polyphyletic or paraphyletic genera disconfirm the generic concepts and you can with confidence hypothesize that the genera should be systematically revised.
Hope that helps! Cheerio!
Maybe. You'd need to be confident about the alignment being good. Also the genes sampled may not be informative for the group. Before giving up would align using MAFFT and visually check it then fix the irregularities. The program Mesquite is good for alignment checking and fixing or bioedit. Also how did you combine (concatenate) the genes? seqMatrix is good for this. Geneious is also good for both if you have access to it.
I use Geneious and I aligned using MUSCLE. 18S and COI are normally used for nematodes. The alignment looks ok, but the taxonomic description for the genera is very poor, so I wonder if the poorly supported clades mean that there is not a good genus separation.
Hey Lidia! Low bootstrap values indicate that there is conflicting signal or little signal in the data set. This may be a problem in the alignment, as Chris suggested. In this case it could be due to an erroneous alignment, which often occurs when the sequences aligned are ambiguous or too diverse. Sometimes, different clades had different rates of evolution (more or less mutation when compared to each other). Also, the absence of intermediate taxa could be a problem. Alternatively, information may be simply absent or too low in your alignment. That is the case when a very conserved DNA fragment has been used and sequences are too similar to each other. Normally you cannot conclude much else from low bootstrap values.
Normally, if a node in the dendrogram is not supported (usually bootstrap below 50) you cannot trust the bifurcations and that is why usually these nodes are collapsed in a tree graph to form a polytomy. If a node is poorly supported (50 to anything around 85 - a bit a case of personal interpretation, I think) it is to be treated with care but there is some support at least. Anything above that value can be interpreted with confidence.
I guess that your COI tree will have problems predominantly. Here, due to the protein-coding nature of this gene, the alignment should not be a problem. However, the third codon position may lead to "background noise" in your data because of the saturated code and silent mutations. A way to improve your results may be to infer relationships again with the third codon positions excluded.
As for the robustness of the concepts of the genera you study: you may doubt it is, as you wrote, they rely on poor descriptions. Think about the apomorphic morphological characters: are they complex characters, or is there evidence that they may have evolved multiple times independently? Usually, a taxon defined by more and more complex characters is more robust. I would not accept my molecular results blindly, but investigate, if they are plausible given alternative data (morphology). If so, robustly supported clades in your dendrogram that show polyphyletic or paraphyletic genera disconfirm the generic concepts and you can with confidence hypothesize that the genera should be systematically revised.
Hope that helps! Cheerio!
Hi Lidia,
If your alignment is good and you don't have any iregularities than there are several options. It is perfectly possible that there is not a good genus seperation at the molecular level and it seems the most easy solution. However you should have look at the region of the 18S and COI that you have amplified. Genetic diversity of 18S may be insufficient to separate genera. COI should have sufficient resolution to separate you genera, taken my knowledge of COI in nematodes into account. However, if you are convinced that the four genera are truely four genera (based on other features) than there is always the possibility that certain (common) environmental features might select against the amplified COI fragment, giving you a smaller difference between species. Check also the pairwise distances between genes from the supposedly different genera, as well as the addition of a molecular clock.
Cheers
- Badges
- Science method
- Similar topics
- Biological Science
- Evolutionary Biology
- Phylogenetics
- Phylogenetic Analysis