search

glennhickey / sg2vg

Prototype code for importing a Side Graph (from a GA4GH server) into a VG
1 stars 2 forks source link

Support new vg is_reverse format #10

Open adamnovak opened 9 years ago

adamnovak commented 9 years ago

@ekg recently moved the is_reverse flag out of Mapping and into Position. This changes the way sg2vg will need to spell paths that go through nodes backward.

glennhickey commented 9 years ago

Yep, this will touch vg2sg as well.

On Mon, Nov 16, 2015 at 2:07 PM, adamnovak notifications@github.com wrote:

@ekg https://github.com/ekg recently moved the is_reverse flag out of Mapping and into Position. This changes the way sg2vg will need to spell paths that go through nodes backward.

— Reply to this email directly or view it on GitHub https://github.com/glennhickey/sg2vg/issues/10.

adamnovak commented 9 years ago

As per discussion on the call, we're going to postpone this until after we finish the reads bake-off.

I'm thinking right now we should target b0e29864e571f2d1ebd7c7600bcebbccfee5224d.

glennhickey commented 9 years ago

That version looks okay by me.

On Mon, Nov 16, 2015 at 3:00 PM, adamnovak notifications@github.com wrote:

As per discussion on the call, we're going to postpone this until after we finish the reads bake-off.

I'm thinking right now we should target b0e29864e571f2d1ebd7c7600bcebbccfee5224d.

— Reply to this email directly or view it on GitHub https://github.com/glennhickey/sg2vg/issues/10#issuecomment-157153391.

ekg commented 9 years ago

Will you ignore path annotations in the bakeoff? On Nov 16, 2015 10:26 PM, "Glenn Hickey" notifications@github.com wrote:

That version looks okay by me.

On Mon, Nov 16, 2015 at 3:00 PM, adamnovak notifications@github.com wrote:

As per discussion on the call, we're going to postpone this until after we finish the reads bake-off.

I'm thinking right now we should target b0e29864e571f2d1ebd7c7600bcebbccfee5224d.

— Reply to this email directly or view it on GitHub https://github.com/glennhickey/sg2vg/issues/10#issuecomment-157153391.

— Reply to this email directly or view it on GitHub https://github.com/glennhickey/sg2vg/issues/10#issuecomment-157175642.

glennhickey commented 9 years ago

Hey Erik,

We're using paths from the submitted graphs if they are available. They're needed for a couple of Sean's basic benchmarks. Adam's new grpah comparison code, which we're still integrating, will also make use of paths.

Apart from that, we're not using paths for alignment or variant calling (except for my attempts to call on surjections to the reference, which I'm not sure will go anywhere).

On the vg end, we're basing this on code that assumes the old format, which is one of the reasons not to change right now (despite looking forward to testing out the new build)

cheers -Glenn

On Mon, Nov 16, 2015 at 4:30 PM, Erik Garrison notifications@github.com wrote:

Will you ignore path annotations in the bakeoff? On Nov 16, 2015 10:26 PM, "Glenn Hickey" notifications@github.com wrote:

That version looks okay by me.

On Mon, Nov 16, 2015 at 3:00 PM, adamnovak notifications@github.com wrote:

As per discussion on the call, we're going to postpone this until after we finish the reads bake-off.

I'm thinking right now we should target b0e29864e571f2d1ebd7c7600bcebbccfee5224d.

— Reply to this email directly or view it on GitHub <https://github.com/glennhickey/sg2vg/issues/10#issuecomment-157153391 .

— Reply to this email directly or view it on GitHub https://github.com/glennhickey/sg2vg/issues/10#issuecomment-157175642.

— Reply to this email directly or view it on GitHub https://github.com/glennhickey/sg2vg/issues/10#issuecomment-157176542.

ekg commented 9 years ago

Thanks for the info! On Nov 16, 2015 10:44 PM, "Glenn Hickey" notifications@github.com wrote:

Hey Erik,

We're using paths from the submitted graphs if they are available. They're needed for a couple of Sean's basic benchmarks. Adam's new grpah comparison code, which we're still integrating, will also make use of paths.

Apart from that, we're not using paths for alignment or variant calling (except for my attempts to call on surjections to the reference, which I'm not sure will go anywhere).

On the vg end, we're basing this on code that assumes the old format, which is one of the reasons not to change right now (despite looking forward to testing out the new build)

cheers -Glenn

On Mon, Nov 16, 2015 at 4:30 PM, Erik Garrison notifications@github.com wrote:

Will you ignore path annotations in the bakeoff? On Nov 16, 2015 10:26 PM, "Glenn Hickey" notifications@github.com wrote:

That version looks okay by me.

On Mon, Nov 16, 2015 at 3:00 PM, adamnovak notifications@github.com wrote:

As per discussion on the call, we're going to postpone this until after we finish the reads bake-off.

I'm thinking right now we should target b0e29864e571f2d1ebd7c7600bcebbccfee5224d.

— Reply to this email directly or view it on GitHub < https://github.com/glennhickey/sg2vg/issues/10#issuecomment-157153391 .

— Reply to this email directly or view it on GitHub <https://github.com/glennhickey/sg2vg/issues/10#issuecomment-157175642 .

— Reply to this email directly or view it on GitHub https://github.com/glennhickey/sg2vg/issues/10#issuecomment-157176542.

— Reply to this email directly or view it on GitHub https://github.com/glennhickey/sg2vg/issues/10#issuecomment-157180080.