Examples of using `vg view` / generating dot files about vg

[ekg]

Here's how you output dot format:

vg view -d x.vg >x.dot

Every tool in vg is self-documenting. Just add -h or --help to the command line.

Also, any graph output can be piped into any other program. So you can pick up a tiny bit of one graph and then pipe it into vg view or vg align. For example, here we can index a graph we've made, then pick up the node with id 50 and its neighbors (defined by a BFS search with up to 5 steps):

vg index -s x.vg
vg find -n 50 -c 5 x.vg | vg view -d - | dot -Tpng -o barrio50.png

These should be added to the documentation. My preference is to start documenting things on the wiki.

[ryan-williams]

cool, thanks @ekg. I'm happy to add some documentation things as I play around with vg a bit; are you suggesting creating this repo's GitHub wiki? Do you envision the README remaining a/the main source of documentation?

Also, while we're here, any interest in making some of the unchecked boxes from the README into issues?

[ekg]

I think we should start the wiki. The readme should be shorter and more
clear so people don't get confused.

And the unchecked boxes should absolutely be issues. It hasn't been an..
Issue because its been just me for the past four months. Pick your favorite
and let's jam.
On Mar 13, 2015 8:46 PM, "Ryan Williams" [email protected] wrote:

cool, thanks @ekg https://github.com/ekg. I'm happy to add some
documentation things as I play around with vg a bit; are you suggesting
creating this repo's GitHub wiki? Do you envision the README remaining
a/the main source of documentation?

Also, while we're here, any interest in making some of the unchecked boxes
from the README into issues?

—
Reply to this email directly or view it on GitHub
#7 (comment).

[ekg]

Have you had a chance to experiment further?

[ryan-williams]

Sorry for the radio silence, was traveling and otherwise preoccupied but have time to work on variant-graph-related things now. Beginning the "empty checkbox" -> issue migration now, wave me off if that's not a good idea!

[ekg]

No, it's great! Go for it. Will you be interested in working on development?
On Apr 14, 2015 8:06 PM, "Ryan Williams" [email protected] wrote:

Sorry for the radio silence, was traveling and otherwise preoccupied but
have time to work on variant-graph-related things now. Beginning the "empty
checkbox" -> issue migration now, wave me off if that's not a good idea!

—
Reply to this email directly or view it on GitHub
#7 (comment).

[ryan-williams]

Interested for sure; not sure I should commit to anything yet but if you have suggestions for pieces I could bite off easily I am all ears. I have yet to use vg beyond running it on its own test cases and trying to follow the GA4GH discussion around sub-graphs vs. side-graphs.

I've been thinking a lot (and coding a little) about what a good, interactive, web-based graph-viz story would look like. Can take that discussion over to #8 though.

[ekg]

The tests are there so you can see if any commits you make will break the system, so don't be afraid to try out new ideas. What are you interested in working on? There are obviously a number of open tasks in vg. The major push in the next few months will be to use a GCSA derivative for decreasing the size of the sequence index. I see this as critical for making adoption of this aligner tenable Other than this, I'd like to deal with circular genomes, to use an alignment to modify the graph (allowing iterative construction for alignable genomes), generate a compressed version of a sequencing dataset, and add annotations to the graph. These seem like essential features, and if you just want to get your hands dirty any will suit. I am begging the GA4GH to consider the models that vg is using as a potential interchange definition. I don't know how successful I'll be, but I can ensure that vg is a working, practical, and well-known system. On the standardization front, the thing to do is to provide a pull request adjusting the GA4GH model to be congruent with vg (variation graphs, which are sequence graphs with labeled paths). I am increasingly reticent to work on that effort because there is not another working graph aligner/genome implementation in the public, but hopefully this will change soon. I'll help you work on anything you'd like to contribute. Issues are a great and public place to discuss things, but if you need I can be reached in realtime as well. Just email me and we can set up a time to discuss.

[ryan-williams]

That sounds good @ekg; I don't see your email here but mine is [email protected] if you want to shoot me a note; setting up a call or something sounds great.