Notebook to service

[david4096]

Like the count everything/beacon, it is in principle possible to build up our service architecture using lightweight clients over the GA4GH API. This script demonstrates ExAC and GA4GH being connected to answer a query GA4GH alone couldn't answer.

In principle, any analysis to be turned into a service. This allows analyses to be published as services that can be pointed at different data, as opposed to programs that need to be run. In principle, any python notebook could be translated into a service.

[diekhans]

Can you please elaborate, or link to context? I don't understand what this means (although it sounds cool).

David Steinberg notifications@github.com writes:

Like the count everything/beacon, it is in principle possible to build up our service architecture using lightweight clients over the GA4GH API.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.*

[Kusdhill]

I’m currently working on a notebook where users will be able to search a gene for specific type(s) of transcript effects, and these search results will then be compared to data in the 1000 Genomes data set. Here is a good example of what the output will look like:

HG02010 has missense_variant at position 41234480 to 41234481 HG03753 has synonymous_variant at position 41234505 to 41234506 etc.

One idea I had that would make this output easier to examine is to place the output in an HTML file and use Flask with jQuery to allow users to mouseover things like “HG02010” to see more information about that particular person. Color coding the transcript effects based on the seriousness of variations could also be very helpful. Overall, I think this could go a long way in demonstrating how the API can be used to create new services.

Here is my current work flow:

• fix current model of researchingRegions (this is a rudimentary version of the program I described above)
• replace start/stop parameters with gene search
• get started with Flask and jQuery
• integrate/translate working python notebook into Flask/jQuery

[david4096]

We should consider that some queries won't be able to be handled synchronously, so for more complicated queries we may need to give a result ticket which can be checked for completion.

[diekhans]

Oh, I was right, this is cool! The mouse over for more information is really useful. The Ensembl genome browser makes good use of it (well, you have to click).

Kunal Dhillon notifications@github.com writes:

I’m currently working on a notebook where users will be able to search a gene for specific type(s) of transcript effects, and these search results will then be compared to data in the 1000 Genomes data set. Here is a good example of what the output will look like:

HG02010 has missense_variant at position 41234480 to 41234481
HG03753 has synonymous_variant at position 41234505 to 41234506
etc.

One idea I had that would make this output easier to examine is to place the output in an HTML file and use Flask with jQuery to allow users to mouseover things like “HG02010” to see more information about that particular person. Color coding the transcript effects based on the seriousness of variations could also be very helpful. Overall, I think this could go a long way in demonstrating how the API can be used to create new services.

Here is my current work flow:

• fix current model of researchingRegions (this is a rudimentary version of the program I described above) • replace start/stop parameters with gene search • get started with Flask and jQuery • integrate/translate working python notebook into Flask/jQuery

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.*