How to conclude two genome sequences have the same origin?

Question

[Crunk]

Hi, I have the following problem:

I have two genome sequences of two phages and they are of equal length, the result is 100% similar when blasting the two sequences (but there are about 15 different regions, each region length ranges from 50-100 bps) . But the microbiological results show that they have plaques of different sizes.

I wonder what conditions I need to confirm that the two sequences above come from the same phage.

[Ram]

Why are you quoting yourself instead of directly asking the question?

[Crunk]

Oh, I just wanted to highlight the problem I had. I am a newbie. Sorry for making you uncomfortable

[Ram]

It's just a unique way of expressing your problem, not really a question of my comfort. Welcome to Biostars!

[Joe]

Presumably by "same phage" you mean the same 'species' (or whatever the equivilant taxonomic rank for your phage is)?

You would be best off trying to solve this phylogenetically. Take a selection of known phage genomes that are more distantly related, and a reference sequence for this specific phage (if one is available) and see how closely they group together on a tree.

I don't know how big your genome is but ~1500 bp of total divergence is likely to be not much at all so you're be pretty justified in saying they're the same phage I think, but as others point out, don't be fooled by local identity.

Answer 1

[Carlo Yague]

Local alignment of large sequences will always find short regions with the highest homology similarity (see comment by @lieven.sterck) possible (hence, the 100% similarity in BLAST), but to know for sure that the two genome come from the same organism, you need to do a global alignment. For instance you can use stretcher, I think it should work well with a viral genome.

Difference between global and local alignments:

[lieven.sterck]

Hey Carlo,

sorry, sensitive point, but

highest homology possible

that gets me all wind up :-)

homology is a binary measure so there are no degrees nor more or less homology.

I assume it should read similarity .

otherwise your answer makes perfect sense addressing the issue here!

[Carlo Yague]

You are, of course, correct (I knew this, but it is always to wrong word to comes first to mind if I'm not careful). I have edited my post accordingly.

[lieven.sterck]

no worries ;)

Answer 2

[BioinformaticBird]

I guess you can not do it with only two genomes. Consider at least three genomes to make an estimation about the origin. Let's put it in the other way: You can always compare two genomes to each other and will find mostly some changes, independently of their relatedness. How should you be able to make any guess based on only this information?