I have aligned protein sequences using Clustal W but how do I confirm/validate whether this alignment is correct? How can one validate a phylogenetic tree?
Hiya,
Another option is to calculate the norMD score (http://www.ncbi.nlm.nih.gov/pubmed/11734009), which gives you a quality score of the alignment. >0.60 indicates a good alignment.
Depending on what software you use, you will end up with a slightly different alignment. You might like to try a few different alignment softwares (PRANK, Muscle, MAFFT) and either (1) draw trees from all alignments or (2) calculate the norMD score for each of the alignments and draw a phylogenetic tree from the best scoring alignment.
Otherwise you could explore BaliPhy which is software that explores all possible alignments and tree space in MCMC framework. This method would be the most computationally intensive, but in theory should give you the most accurate answer.
I hope this helps.
You can take a look at 'Phylogenetic trees made easy', Barry G. Hall. According to this, (1) as long as the alignment is >50% accurate, increasing accuracy of the alignment has little effect on the accuracy of the tree. The suitability of the alignment for phylogenetic analysis can be checked by looking at the average amino acid identity. If the average amino acid identity is > 30%, then the alignment is accurate enough. (2) Moreover, if the Jukes-Cantor distance is > 1, you cannot use the data to make Neighbor-Joining trees and have to use another method.
Moreover isn't jukes-cantor method for nucleic acids? I am dealng with protein sequences..
Hiya,
Another option is to calculate the norMD score (http://www.ncbi.nlm.nih.gov/pubmed/11734009), which gives you a quality score of the alignment. >0.60 indicates a good alignment.
Depending on what software you use, you will end up with a slightly different alignment. You might like to try a few different alignment softwares (PRANK, Muscle, MAFFT) and either (1) draw trees from all alignments or (2) calculate the norMD score for each of the alignments and draw a phylogenetic tree from the best scoring alignment.
Otherwise you could explore BaliPhy which is software that explores all possible alignments and tree space in MCMC framework. This method would be the most computationally intensive, but in theory should give you the most accurate answer.
I hope this helps.
Bootstrap value is important to evaluate the robustness of the branch, if you have high bootstrap values (higher than 95%) you are confident with your tree (the topology at the branch is "correct"). Bootstrap consensus tree can be easily performed with MEGA software
You can also try TCoffee for your alignment and sometimes you would need to trim that alignment file (from TCoffee) with GBlock (if the target region is too variable) before any phylogenetic analyses.
Gblock helps in excluding ambiguous gaps and reduces the long branch effect in the phylogenetic tree.
Bootstrap (ML tree) and posterior probabilities (bayes tree) are important, node values : > 70% (bootstrap) and > 95% PP (bayes) are good (robust)
For alignments, MCoffee combines several different alignment methods, and shows the alignment colour-coded according to how well the different alignment methods match, so a red region means the different methods agree and the alignmnet is likely to be robust, whereas a blue (if I remember correctly) region means different methods produce different alignments so the alignment is ambiguous.
For phylogenetic trees, use bootstrap values or Bayes posterior probabilities, as Romina C says, to indicate confidence that the tree obtained is the optimal tree for your sequence data under the model and optimality criteria used (but not necessarily that the tree is a correct evolutionary history as this will depend on the assumptions of your method, homoplasy in your chosen genes, etc).
For validating multiple sequences and identify specific races you can use NTsys Sofware.
- Badges
- Science method
- Similar topics
- Biological Science
- Evolutionary Biology
- Phylogenetics
- Phylogenetic Analysis