reverse complement about dnaio HOT 6 CLOSED

I agree. I wanted to offer a quick fix that could be used now, and then failed to answer the original question. Sorry.

Note that a proper function should also work for wildcards and lowercase characters (and the example you gave is missing the "reversing" part.)

And then gave a bad answer too by not reversing... It's not a great day today.

I wouldn’t mind moving this to dnaio. Just as we have methods for slicing a SequenceRecord, computing a reverse complement is also a thing one regularly needs to do with DNA sequences, so it’s IMO natural to offer this.

I see cutadapt uses a translate table then reverses using Python. This is fast enough, but there are two caveats

• Python uses a dict to translate, because there are so many unicode code points. Since we guarantee 1-byte kind strings we can use a 256 character table. (bytes.maketrans automatically creates one of these as well).
• We can do the reversing and translation at the same time. Creating a new unicode object and then populating it with the translated nucleotides in reverse.

When we do that in Cython it should be quite a bit faster than the python method so this should also be a win for cutadapt.

SequenceRecord and BytesSequenceRecord can share the underlying method easily I think. Just as with the is_mate function.

from dnaio.

It is python builtin:

from dnaio import Sequence

complement_table = str.maketrans(dict(A="T", C="G", G="C", T="A"))
my_sequence = Sequence("my_seq", "GATTACA", "HHHHHHH")
my_sequence.sequence.translate(complement_table)

Results in 'CTAATGT'

Does this work for you?

from dnaio.

Reverse complements aren’t hard, but there are things one can get wrong, so it’s worth having a method for it. I would definitely not call it "built in". Note that a proper function should also work for wildcards and lowercase characters (and the example you gave is missing the "reversing" part.)

Here is what I use in Cutadapt:
https://github.com/marcelm/cutadapt/blob/26d3f39b6a37bcee8383d7d8f4a95879528137e4/src/cutadapt/utils.py#L172-L189

I wouldn’t mind moving this to dnaio. Just as we have methods for slicing a SequenceRecord, computing a reverse complement is also a thing one regularly needs to do with DNA sequences, so it’s IMO natural to offer this.

from dnaio.

Oh great, it sounded as if you were opposed to adding the functionality. And I’m of course all for speeding things up!

from dnaio.

Hi all,

Thanks so much for the comments. I'll have a go with str.maketrans and translate

Best
Jorge

from dnaio.

Thanks so much for the comments. I'll have a go with str.maketrans and translate

And for the reverse part there is a built-in reversed function: https://docs.python.org/3/library/functions.html#reversed . Good luck!

In the meantime we may hammer out a method, but I can give no ETA at this point. My time is rather limited now.

from dnaio.