BLAST n not similar to BLASTx - (Apr/16/2007 )

Dear all,

What happens when a gene sequence I BLASTed in BLASTn gave me Gene A cds as the highest alignment similarity (179/171) (Strand Minus/Plus) and Gene B (150/160) (Plus/Plus) but the same gene sequence in BLASTx gave me Gene B as the highest alignment similarity. So should I accept Gene A as my sequence homolog or Gene B?

Many thanks,
Chris

---

En, quite interesting, here is my thought:
1. blastn is used to blast nucleotide sequences (hence n), whereas blastx (use nucleotide to blast protein, i.e. amino acid sequences), hence x.
2. you also want to see the length of the match, if they are in coding sequence.
3. When you say gene A, and gene B, are they really two different gene, or same gene, difference sequences (different clone, reference sequence, protein sequence vs nucleotide sequence)? Frankly, I don't see any difference here, the similarity is very convincing, I would say both of them are right.

---

AFAIK, the only reason this could happen is when the nucleotide alignment (blastn) is better for gene A than B, and the amino acid alignment (blastx) is better for gene B than A. The scores seem quite good, which suggests there may be a frameshift in gene A which disrupts the amino acid sequence. For instance
a: ATGCAAGAACCA
x: ATG-AAGAACCA
is a good hit, but
a: Met Gln Glu Pro
x: Met Lys Asn *
is not a good hit.

Try blastx with the -w switch for frameshifts

---