Thanks! So is it a good idea to use BLASTN for non-coding regions and for coding regions, to use TBLASTX (nucleotide vs. nucleotide, both translated in all 6 frames)?
I've read in many papers that the human non-coding RNA Xist (involved in X chromosome inactivation) is "poorly conserved" in other mammals because it only has 60% homology.
Isn't 60% conservation a lot?
So I thought maybe that's not a lot compared to protein coding genes that are typically very well conserved BUT:
- I ran BLASTN on the human vs. cow Xist (result here) and I get 59% query coverage.
- I ran BLASTN on the human vs. mouse TP53 coding gene (result here) but I get 24% query coverage.
I must be doing something wrong... Do I need to remove introns before doing the protein-coding query? Is the query coverage completely unrelated to the conservation score? Should I not be using BLASTN for conservation calculations?
1 answer
60% is not a lot, especially when you consider that many proteins are 90-100% conserved at the amino acid level. Your mistake is twofold:
1) You're including non-coding parts of the genes (introns), where there is much less functional constraint.
2) You are comparing at the DNA level instead of the amino acid level. A DNA change from CCU to CCC still gives you proline, so there's typically no selection for or against that change.
Log in to answer this question.