GenSpectrum / cov-spectrum-website

A web platform to detect and analyze variants of SARS-CoV-2
https://cov-spectrum.org
GNU General Public License v3.0
61 stars 14 forks source link

Add VILOCA results to dashboard #813

Open mcarrara-bioinfo opened 1 year ago

mcarrara-bioinfo commented 1 year ago

As initially discussed through other channels, CBG developed VILOCA to detect potential drug-resistance mutations in wastewater data and they wish is to add the results of this analysis to the cov-spectrum dashboard.

We have compiled a document with comments, the text that we wish to add and the figure to include. I also attached to this issue an example of the actual file to be shared and embedded. The output of the VILOCA run will be directly the figure to embed in the page.

The image is to be updated every time a VILOCA run is complete, which is likely to be with a slower frequency and at different time frames than the updates to the Swiss curves.

A discussion is necessary for the preferred method to exchange the figure.

mutation_calls_on_potential_drug_resistance_sites.pdf

corneliusroemer commented 1 year ago

The Google doc seems to not allow link-only viewing, is that on purpose?

chaoran-chen commented 1 year ago

Hi @mcarrara-bioinfo, I assume that it wouldn't be a problem to provide the plot as an SVG? I think that an SVG will be easier to embed on a website.

Should we maybe just create a new database table where you can store the content of the SVG? This should be the easiest because it is an already established method to exchange data.

mcarrara-bioinfo commented 1 year ago

The Google doc seems to not allow link-only viewing, is that on purpose?

I have changed the access permissions, you should be able to access now.

Hi @mcarrara-bioinfo, I assume that it wouldn't be a problem to provide the plot as an SVG? I think that an SVG will be easier to embed on a website.

I am confirming this with the developer, but I don't think that would be a problem.

Should we maybe just create a new database table where you can store the content of the SVG? This should be the easiest because it is an already established method to exchange data.

On our side there is no preference on the exchange method as long as it's feasible from machines within the BSSE network and from Euler.

chaoran-chen commented 1 year ago

I just created the table wastewater_viloca which has a single column plot_svg (type: TEXT). Would that work? You should already have access to it.

chaoran-chen commented 1 year ago

The table has exactly one row (which currently contains "nothing here yet, please update!") and you can update it.

mcarrara-bioinfo commented 1 year ago

@chaoran-chen I have confirmed that the tool can be adapted to generate SVG files. I see the table and we can include code to push the SVG whenever new results are available.

I am currently working on the final tests, then it's a matter of waiting for a PR review on the CBG automation repository.

mcarrara-bioinfo commented 1 year ago

The tool has been adapted to generate SVG files. I attached an example. We will start filling and updating the new table as soon as the automation that is supposed to run VILOCA is updated (code currently under review). At that point the heatmap can be published.

Question: in the current situation, will the heatmap be published automatically as soon as I fill the table or do you need to finalise something in the frontend?

mutation_calls_on_potential_drug_resistance_sites

chaoran-chen commented 1 year ago

No, we need to implement the integration into the frontend first. Could you please deposit a version of the chart so that we have something to work with?

mcarrara-bioinfo commented 1 year ago

I uploaded to the database a test heatmap you can use as reference

mcarrara-bioinfo commented 1 year ago

Delayed because of a much-needed refactoring of the automation that is supposed to generate the data and because of my holidays. I will resume working on it next week. I estimate everything to be ready to deploy by the beginning of October, considering testing and ironing out details after deploying the changes on our side. I will check if I can start a non-automated run to build the most recent results, so that you can merge and close the issue earlier by publishing it while I finish with the automation. I will update you on this last point next week