sequencing/NxTrim

specify mp/unknown only with `--stdout`

jaredo opened this issue · 9 comments

We need to be able to pipe definite MP reads to stdout for some applications. For example, piping them directly to bwa for alignment prior to scaffolding.

As discussed in #18

What do you think about adding a switch to decide whether --stdout writes just mp or a "combined stream" with mp/unknown (as it does atm)?

Yes, this is what I am thinking. I will add flags like:

--mp: only output mate-pair reads
--unknown: only output unknown reads

(or something to that effect)

... which may work combined?
--mp --unknown

Yes, but I think mp+unknown would be the default behaviour. Some care is required here though. Do I include a --pe flag? And --se is a no-go for streaming because it will break the pairing.

More generally, I need to re-work the command line arguments a bit because they have become unclear. NxTrim started as a very specific use-case - frugal trimming of reads to support assemble from relatively low coverage data. Extra functionality has been tacked on post-hoc.

mp+unknown as default, yes, that's probably better.

--pe as additional flag? Why not, could be helpful in some situations.

NxTrim is doing a great job, thanks and thumbs up for your work!

I am glad it is useful. I think I can take a look at this next week.

I have just added the options --stdout-mp and --stdout-un. I believe the former will do what you want for scaffolding (pipe the guaranteed mate-pairs to stdout).

And --stdout dumps unknown+mp?

Great, thanks :-)

And --stdout dumps unknown+mp?

Yes exactly.