Hi folks, I am comparing two rather divergent homologs of the same protein-coding gene and am interested in whether there has been any regional variation in selection along its coding sequence. I used DnaSP to perform a sliding window analysis of their Ka/Ks ratios. The proteins are only 40% identical on the amino acid level, which a reviewer has suggested is too divergent for these kinds of analyses. I am inclined to believe that he/she is correct, but I can't find any citation to support this. Does anyone know of one, or have personal experience in these types of analyses? Much thanks.
I believe that one of the reasons why proteins with high divergence levels are bad for Ka/Ks analysis is because it is likely that all possible synonymous sites have already mutated (maybe even multiple times). If the number of matching synonymous positions shared by the two homologs is similar to the one expected by chance you are most likely underestimating Ks and thus possibly overestimating Ka/Ks (false positive positive selection?). However, the reviewer might have had something else in mind. There was also a method published that detects past positive selection from ongoing negative selection, but it might require additional data, here is a link, I hope it helps: http://www.ncbi.nlm.nih.gov/pubmed/21859804
http://www.ncbi.nlm.nih.gov/pubmed/21859804
On the basis of Kyte Doolite analysis a protein which is said to be marker protein must have homology more than 40 per cent, but proteins with high divergence levels show bad for Ka/Ks value or ration, It display that all possible synonymous sites have already mutated and no further chnage is possible more than 10-30%, a protein with a single amino acid change may be selected some time for a free going mutation under nature, but in comparison if some want single protein to be selected among other proteins of high homology should need differences in active site regions, it these more than 5% it will be selected automatically, but below this selection will operate by chance.
Homologs are not supposed to compare for positive selection, rather orthologs are required to do the analysis.
So: you have to
1. Find similarity in rRNA to see if the species belongs to same clade
2. identify the proteins as orthologs using reciprocal blast (generally e-10)
3. Remove the paralogs and xenologs from the datasets
If all these process indicates that your sequences are orthologs then you can ignor how much they are identical in protein level.
As positive selection is a measurement of nucleotide substitution therefore amino acid smilarity is not the initial factor..... (they are only considered at the final stage of the analysis)
- Badges
- Science topic
- Similar topics
- Biological Science
- Evolutionary Biology