Future plans
Kristian Gray
Internal database
- In 2014 we announced that the PostgreSQL database server should be replaced by a MySQL server.
- Since we needed to replace drupal and redesign/design the HGNC and VGNC websites this had to be delayed.
- Within the next funding period we would like to revisit this issue.
Internal database - problems
- Current schema limitations make development difficult
- Parts of the schema no longer reflect the biological data
- Database schema is not truly relational
Internal database - solutions
- Use MySQL which is supported and is the same database as our front end
- Redesign the schema to better reflect the data and fully normalise
- This will allow us to add more sources of data and allow us to query the database easier for reports and development
- The VGNC is already using a similar database schema which was originally created as a prototype for a new MySQL HGNC internal database.
Internal applications and data release pipeline
- New database = new applications + new data release pipeline
- Rewrite application using modern and widely used frameworks and layouts (AngularJS with bootstrap CSS etc)
- Change the update and release pipeline so that data downloads happen in parallel (i.e. using the farm).
Release the new HGNC beta site and retire the old
- Need to finish of the beta site
- - Adding xrefs to RNA central to the symbol reports
- - Multiple Ensembl ENSG IDs to the symbol reports
- - creating a tag at the top of the symbol reports for families
- Need to do extensive user experience testing
- Need to write redirect rules on the server.
- Complete the website editor.
- Make all our resources HTTPS accessible.
REST API
- The HGNC REST API has been a successful addition to our service.
- In one year we have had 35,500 sessions with 81% of the sessions created by returning users.
- Because of this we would like to expand our REST API to include gene families and the VGNC data.
- Also we could create a web application that allows users to build queries that utilise the REST API in the background.
- The new query builder could then replace the old custom downloads tool.
Infrastructure changes
- Every 5 years the tender on the external data centre (DC) is renewed (2 years into the 5).
- Unlikely that the current data centre contract will be renewed.
- Although a change in DC shouldn't affect the infrastructure model, the reality is that with each DC move, we have had to work on the model.
- We have so far moved from physical machines -> VMs -> microservices and containers ??