Open almosnow opened 1 month ago
Currently the map shows biosamples that match what the user searched for, which could be virus related or other metadata
I think the contigs will tie in if/when we integrate the app with the kmindex which is in development. Then user's can enter any sequence and kmindex would return contigs and SRA accessions with matching kmers and those can be used as the identifiers
As per the Virome data-model, the unit of data which is being represented in the interface is individual contig
which are grouped into virus sOTU
See: https://github.com/serratus-bio/open-virome/wiki/Virome
To display the geospatial distribution, it makes sense to represent each contig
as a point on the graph, which will allow for things such as "Virus Identity Filtering" or "Virus class filtering" and better capture the density of virus-observation per unit geography.
This is going to require a JOIN
operation most likely on the datasets which can be rendered (and their lat/lon) and the data which is already calculated in the virome modules.
Please add a Time Estimate:
for this issue.
This is a small sub-project on its own.
@ababaian you want me to draft all the subtasks?
They're more or less:
That's the right level of analysis yes, each of those sub-tasks should be given a time estimate and that will provide us with an overall time estimate for this task. This is exactly the information which we would want to have thought out upfront when opening such an issue so that it can allow for planning and execution of this in an efficient manner. Let's indeed use this as an example issue for planning out the logistics; feel free to edit the main issue post here and provide the plan for execution.
... because we ultimately want a map of the molecular evidence of an occurrence of a virus, not a map of the experiments that were performed