Keynote: Solving Molecular Analysis of Cancer: From DNA to Proteins
summary
Precision oncology requires moving massive real-time pipelines: terabits per second networks carry 50x whole‑genome slices that produce ~150 GB BAM or ~90 GB CRAM per patient and only ~3 million variants from 3.3 billion bases. The stack uses Kafka and Knife for lossless data flows, Cassandra and Spark for storage and in‑memory analysis, FPGAs (the "dragon" chip) cut a 4 hour step to under an hour, and quantum and blockchain are proposed; current data without germline causes 69% false positives and 26% false negatives.
AI-written summary · about
chapters (13)
- 0:00Intro
- 1:37Precision Medicine
- 3:10Genomic Data
- 5:03EvidenceBased Information
- 7:44Sequencing
- 11:30Data Pipeline
- 12:24FPGAs
- 15:05The Problem
- 15:49Logic
- 17:51Technology
- 18:50Data Storage
- 19:45Open SDN
- 23:13Privacy
related talks
08mKeynote: The Challenge and Opportunity to Build a Transparent Network
22mResearch SuperCluster(RSC): Accelerating AI Research at Scale
20mKeynote
23mNetworking for GenAI Training and Inference Clusters
15mLive Q&A with Speakers – Joseph Provine
16mThe Next 10 Years - Part 1
24mMission Possible: Forging a New Paradigm in 5G Deployment...
22mdShark: A General, Easy to Program and Scalable Framework for Analyzing In-network