Improved version of GWAS Diagram with fast loading and new features

[ljwh2]

As a user, I want to be able to visualise all the associations in the Catalog on the GWAS Diagram and apply useful filters, such as trait. I want the Diagram to load fast and reliably in all major browsers. I want to be able to click through from the Diagram to data of interest in the details pages.

Investigation into Daniel’s prototype is documented here: https://app.zenhub.com/workspaces/gwas-59df823c4a6feb3786810391/issues/gh/ebispot/goci/903

Must have
-fast and reliable loading in all major browsers

• All currently implemented features working reliably i.e:
• option to display association data for p values <5x10-8
• scalable image to either show all chromosomes or zoom in on a specific region.
• filtering capability to display data by parent category and by EFO trait, which is fast to apply.
• download capability for filtered or unfiltered images.
• display of association data upon clicking on a dot, and click through from here to the details pages.
• Solution for x/y axis scaling issue for many dots per cytoband

Should have

• option to display all association data regardless of p-value.
• enhanced filter capability. This should include (in order of priority):
• P-value - set the p value threshold for associations to display
• ancestry group - filter by discovery structured ancestry. Some studies have multiple ancestries, could include an option to include or exclude studies where the ancestry is included together with others
• Journal publish date - set filter according to journal publish date, so that users can create their own time lapse images.
• PMID - show associations only for a specific publication
• Sample number (but only if we can accurately represent sample size - see Enhancement to Studies table)
• Combined filters, e.g. trait and ancestry
• the dots on each band ordered by parent Category, at the moment the order is apparently random.

Could have

• additional new parent categories and colours e.g. protein measurement, metabolite measurement, brain measurement. This would need input from UX or graphic design to identify appropriate colours

[ljwh2]

@eks-ebi to flesh out UX design in first instance

[jdhayhurst]

Dev prerequisites:

• update python version
• Add poetry
• update dockerfile
  Development:
• change response model
• integrate with ideogram js
  Optional:
• write response and request models with pydantic
• replace flask with fast-api

[jdhayhurst]

All done: EBISPOT/gwas-diagram#13
Some optimisation/constraints may need to be added to prevent large dataset filtering.
TODO:

• proper static/js integration with the API
• deployment

[jdhayhurst]

Investigate mouse over annotations for histogram view

[sprintell]

@jdhayhurst and @ala-ebi to meet on this, it'll be handed over to Ala