The future's Shiny: Pioneering genomic medicine in R
summary
A Shiny MI portal is built to serve the 100,000 Genomes Project and the National Genomics Medicine Service, scaled to hundreds or thousands of concurrent users by combining Docker horizontal scaling with per-instance vertical scaling, caching and asynchronous R using the future and promises packages. Profiling with profvis and shiny load test finds database and API calls as bottlenecks, so heavy processing moves into SQL, predictable screens prefetch data, and rich interactivity uses HTML widgets like Plotly, SunburstR and VizNetwork to keep latency consistent.
AI-written summary · about
chapters (18)
- 0:00Introduction
- 1:31The 100K Genome Project
- 3:29The science behind it
- 4:05National Genomics Medicine Service
- 4:27The data pipeline
- 5:13The packages
- 6:15Priorities
- 6:55Examples
- 7:37Demo
- 9:37Case Hierarchy
- 10:31Next Steps
- 11:17Scaling
- 12:41Load tests
- 13:23Optimizing code
- 14:16Database API calls
- 15:01Data loading
- 15:55Working with Shiny
- 16:32Questions
related talks
27mReal Time Shiny Applications
18mShiny New Tools for Scaling your Shiny Apps
04mSpeeding Up Plots in R/Shiny
18mWait, that’s Shiny? Building feature-full, user-friendly interactive data explorers
30mSeeing the needle AND the haystack: single-datapoint selection for billion-poin...
45mFHIR server scale and performance challenges
36mProgrammatically Performant
28mHigh Performance Data Visualization for the Web