Show me your data!
summary
Shows a working genome browser built with D3.js, TypeScript, Redux and RxJS, backed by a Python streaming service on S3 that serves FastQ/BAM/VCF chunks. Claims: human genome = 3 billion bases (~750 MB compressed), raw FastQ outputs reach 100–300 GB, aligned BAMs 10s–100s GB; single BAM track holds ~50,000 elements so canvas (not SVG) is used for perf, sometimes rendering tiles on the backend or using WebGL. Describes SVG vs Canvas vs WebGL tradeoffs and a reproducible fastq → BAM → VCF pipeline running on cloud instances (examples: 35 CPUs, 64 GB RAM, multi-day jobs).
AI-written summary · about
chapters (22)
- 0:00Intro
- 0:37I will talk about
- 1:03Alternative history of visualization
- 1:24TSV file
- 2:04The message
- 5:11Lord Varis
- 7:30New York 1849
- 7:51Paris mid 1800s
- 8:11London mid 1800s
- 8:43The name of the doctor was
- 9:35Genomics for SW engineers
- 11:57Difference between DNA and RNA
- 13:03Scanning?
- 13:29Library preparation
- 16:03Alignment
- 17:52Mutations - (G- C)
- 18:23Why are we sequencing genome?
- 20:06Prenatal test & SB Platform
- 23:22BHLHE41 12:26275297 G C
- 24:26Genome Browser
- 28:14Voting in Westeros
- 29:31SVG vs Canvas
related talks
23mDistributed Visualization for Genomic Analysis
04mDBTB INT Alyssa Morrow r
42mGenetics in PureScript Without Knowing Genetics
33mVisualizations using SVG, Canvas, and WebGL in Vue
43mVisualizations using SVG, Canvas, and WebGL in Vue
47mData, Design, Code
27mScalable Data Visualization
41mData, Design, Code