Handle multiple PE libs in IDBA plugin

[levinas]

How to get just the paired end libs in the wasp engine?

This recipe was trying to be conscious of the fact that IDBA needs paired libs to work with:

      (if (has_paired READS)
        (prog
          (define id (idba pp))
          (define assemblies (list id sp vt)))
        (define assemblies (list sp vt)))

But it still fails when a mixed library containing single end reads are passed. Is there a way to do something like the following?

    pe = get_paired READS
    (if (not_empty pe)
         (idba pp)...

[cbun]

Hmm, I thought we fixed this in the idba plugin. This sort of granularity should be handled in the plugin by using the data.paired_files property.

We can implement a method like that in wasp too, if we want to take that route.

I'll look into it, can you give me an example of a failing dataset?

On Mon, Dec 8, 2014, 8:30 AM Fangfang Xia notifications@github.com wrote:

How to get just the paired end libs in the wasp engine?

This recipe was trying to be conscious of the fact that IDBA needs paired libs to work with:

  (if (has_paired READS)
    (prog
      (define id (idba pp))
      (define assemblies (list id sp vt)))
    (define assemblies (list sp vt)))

But it still fails when a mixed library containing single end reads are passed. Is there a way to do something like the following?

pe = get_paired READS
(if (not_empty pe)
     (idba pp)...

— Reply to this email directly or view it on GitHub https://github.com/kbase/assembly/issues/272.

[levinas]

Example: https://github.com/kbase/assembly/blob/master/test/arast.t#L158

You can launch this with: ./arast.t mixed_input_with_smart

Yeah I kind of remember we fixed that, but this input is quite "mixed".

[cbun]

IDBA complains that there is more than 1 paired library.

Assuming we want this to do the right thing, what should the behavior be? We could take the first library and discard the rest? Or maybe concatenate?

[levinas]

Ah I see. Perhaps concatenation, but let's move this out of the current milestone.

On Dec 8, 2014, at 5:31 PM, Christopher Bun notifications@github.com wrote:

IDBA complains that there are more than 1 paired library.

Assuming we want this to do the right thing, what should the behavior be? We could take the first library and discard the rest? Or maybe concatenate?

— Reply to this email directly or view it on GitHub.

[levinas]

IDBA now takes the first PE lib, if there is one, and ignores the rest libs. https://github.com/kbase/assembly/commit/9d423424295de40afe2895465534dbc3e931ea73