Bootstrapping of unaligned sequeces

Question

[arit.mhp]

Hi everyone, I have developed a new alignment free method to generate the phylogenetic tree from the Mitochondrial Genome sequences. I need to check the bootstrap supports of my newly generated trees. But I have doubt how can I bootstrap unaligned sequences and how to compute the final bootstrap scores for each clade of the tree. So in this regard, if you would suggest what should my approach to compute the confidence scores of our proposed method. Thanking you,

[Andrzej Zielezinski]

If your alignment-free approach is based on k-mers you can resample the original k-mers with replacement.

[arit.mhp]

no its based on graphical method, representing a genome sequence graphically and computes the distance matrix.

Answer 1

[kloetzl]

Bootstrapping requires an alignments. However, there are other approaches to compute support values. See the following paper, for example: Support Values for Genome Phylogenies.

[arit.mhp]

Hi kloetzl, thank u for sharing this very interesting paper. I got a different view on bootstrapping. But I have a confusion, as I have developed a method which computes the distance matrix by taking the unaligned sequences as input, how this method can be applicable for my case. Waiting for your suggestions.

[kloetzl]

If your method produces a distance matrix in the PHYLIP file format, you are half way there. Download afra (mentioned in the above paper) and it will give you a tree with "bootstrap" values.

[arit.mhp]

Basically, we have developed a alignment-free method using mitochondrial genome sequences which computes three distance matrices using three different methods. From each distance matrix, we get a tree by applying UPGMA. Final tree is formed applying the consensus based method. Is afra is applicable here?

[kloetzl]

afra should be close to what you want, except it does neighbor-joining, not UPGMA. So you might have to fix that in afra or write your own tool.

[arit.mhp]

Ok... Thanks a lot.....