allow plates and sera to be categorized into groups

[jbloom]

Looking at current GitHubs, I think it would be helpful to allow plates and/or sera to be categorized into groups, such as "pilots_for_determining_library_dilution" etc. That way the data/plots for troubleshooting are not mixed with the ones for actual results.

@anloes @ckikawa, any thoughts on this idea or how to do it before I start to implement?

[anloes]

This is a good idea. One strategy would just be to add a label in the config for each plate, probably called "group", and run aggregate_titers on both all plates and on each of the groups, such that we generate an html with data from all plates, and an independent html for each group.

I am not sure if we need to be able to specify multiple groups for some plates. This would allow you to create independent plots for all individuals in two treatment groups, or separate by age cohort, etc. But unless also implemented at the sera level it might encourage running all of a given group together on the same plate, which I don't generally think is a good idea. I am doing this now manually, I do wonder if this is perhaps better done this way? But perhaps adding a template notebook which allows one to start to make these plots is helpful?

[ckikawa]

Agreed! Rather than adding more keys to each per-plate config, would it be better to instead have some sort of centralized dictionary in the config that defines group keys as consisting of certain plates/sera?

This would be nice for identifying samples as belonging to library optimization (as you pointed out) or as belonging to different serum sample cohorts/studies. Particularly if this could be used to interactively toggle sera on the aggreagte_titers plots. It would be helpful to filter the many sera NT50 plots to just those relevant to library optimization, just those relative to specific cohorts, etc.