Getting command line blastn results for the complementary strand

Question

[langziv]

Hello.

Up to now I worked with command line blastn results without seeing any indication for the results being on the + or - strands of a DNA. Is there a parameter to pass with the blastn command in order to get results for both strands?

Thanks!

Answer 1

[lieven.sterck]

If you don't specify it specifically, the default of blast is to search for both strands.

It will also print in the output if it's plus or minus (in the default output at least it should mention "strand: plus/plus" or such). You can also always look at the alignment coordinates and if those are small -> big it's plus, if it's big -> small it's minus strand

[langziv]

Thanks.

I'm using the outfmt 6 output format and couldn't find in which column this information is given.

What do you mean by alignment coordinates?

Maybe you refer to the web blastn application? I'm using the command line blastn.

[lieven.sterck]

ah, yes it's indeed for the tabular output I mentioned the coordinates thing.

Since in the tab output it does indeed not make any mention from the plus/minus strand you have to look at the alignment coordinates. the coordinates in column 7-8-9-10 indicate the start/stop of the alignments. if there it is outputted as large -> small number it means minus strand, if it's small -> large it denotes plus strand.

For both the web as cmdline it should be the same behaviour.

[langziv]

Thanks.

That raises another question: I just checked my blastn output, and it has only small to large positions. Seems weird. Is that a reasonable output considering the small to large and large to small patterns?

The input is a single-stranded fasta file (as all fasta files I've seen were).

From what I understood, blastn creates the complementary strand so that it can search it.

[lieven.sterck]

that is a bit strange but it can happen depending on the specific use case.

Did you check all coordinates for all alignments? (also both for the query & hit?)

[langziv]

Yes. Same thing. Maybe it has to do with the blastn command.

[lieven.sterck]

neh, I wouldn't think so, rather with the input sequences/database .

You could quickly check this: take a few sequences from your input file and make the reverse complement. Blast those against your DB and see what the output looks like (in best case do it together with the few original sequence so you can see the difference)

[langziv]

Ok. I'll try.

Blast's algorithm is supposed to create complementary strands, right?

[lieven.sterck]

it also searches the reverse complement strand (if that is what you are asking) indeed.

Can't remember currently whether it will do this for the query or the DB (gut feeling says query sequence though)