Challenges of analysing the wheat genome
summary
Claim: wheat is triploid, five times more DNA than a human, and up to 80% repeats, so standard pipelines hit memory and logic limits. Alignment files balloon (17 Gb reference yields ~0.5 Tb compressed alignments), single runs take weeks on cluster, FPGAs cut 3–4 day jobs to ~110 minutes but incur slow transfers, and many tools assume model organisms or hard‑coded limits. Practical fixes include rewriting components, running reference‑to‑self to model false positives, and filtering by quality >200 to keep real variants.
AI-written summary · about
chapters (5)
- 0:00Introduction
- 8:04Wheat genome challenges
- 16:46Read alignment in repeats
- 26:01Computational limits
- 32:05Summary
related talks
28mSome Current Computational Challenges in Bioinformatics
51mSimplicity For Scale- Analyzing 15 Million DNA Samples With Python
14mBig data, AI, the genome, and everything
57mBig Data and New Models Needed to Study DNA Variation in Evolution and Cancer
01h 20mFundamental Algorithms in Deep Sequencing II: Mapping/Alignment
30mAlgorithms for DNA Data Storage
53mApplied Functional Programming in Computational Genomics
27mA Statistical, Reference-Free Algorithm Subsumes Myriad Problems in Genome Science