What is the problem with the initial alignment?
The initial alignment is obviously correct [aln] It can easily be turned into a [tree]
To compute this tree:
Compute the alignment using a clustalw server. Do not forget to request Phylip,Fasta or MSF as an output format! Paste the output alignment in drawtree
On this tree, Human appears more related to the Mouse than to the Xenope. The reason for this wrong phylogeny is that although the proteins are homologous, they are not orthologous. In order to gain a better understanding of what is going on, one solution will be to add new sequences.
Adding in new sequences
This may be done by running a blast against Swissprot. Here is a sample alignment obtained this way: [aln] That alignment gave the following tree [tree] There is still a problem: SCG1_HUMAN, SCG1_RAT and SCG1_MOUSE are not in the right order The reason is that their sequences are too closely related at the protein level. A possible solution would be to use the nucleotide sequence rather than the protein sequence There are more neutral sites in coding nucleotide sequences. These accounts for a faster evolution that renders phylogenetic analysis easier
CONCLUSION The first part of this exercise shows that while blast is useful for making a judgement on the existance of some homology between two sequences, more complicated analysis such as orthology or paralogy require a tree as a support.
|