A Statistical, Reference-Free Algorithm Subsumes Myriad Problems in Genome Science

on YouTube ↗27m

" summary

A unifying probabilistic model and a reference-free algorithm called gnomAD operates on raw FASTQ k-mers, builds contingency tables of upstream/downstream k-mer pairs, and returns closed-form P-values for sample-specific sequence diversification without alignment. It runs orders of magnitude faster, detects Omicron-proximal k-mers from 70 South African patients, finds de novo enrichment in the spike protein, resolves nearly identical paralogs (MYL12) into cell-type-specific isoforms, and reduces downstream alignment burden by three orders of magnitude.

AI-written summary · about

" chapters (16)

1. 0:00Intro
2. 0:32Cancer: 1.8 million new cases in the US
3. 2:34Viral infection: historical and ongoing mortality • How and where are virus' evolving?
4. 3:17Common biological theme?
5. 4:32"Human reference genomes" are necessarily incomplete
6. 5:13Use a reference genome?
7. 7:48References are a guidepost towards a fundamental inferential step: detect differences
8. 9:39Shared biology, shared analysis? Sequence diversification
9. 10:56Sequence diversification: fundamental to many biological problems
10. 12:45Sequence diversification: probabilistic statement?
11. 15:58Existing Methods
12. 16:58NOMAD results snapshot
13. 18:07Unsupervised variant discovery in SARS-CoV-2
14. 19:45NOMAD shows diversification in the spike bypassing reference genomes
15. 20:33NOMAD distinguishes homologs: new biology
16. 22:14Biological applications of NOMAD analysis

" related talks

• 53mApplied Functional Programming in Computational GenomicsJamie Dixon · NDC Oslo 2022
• 01h 20mFundamental Algorithms in Deep Sequencing II: Mapping/AlignmentLior Pachter · Algorithmic Challenges in Genomics Boot Camp
• 35mData Driven "Pathway" Analysis with ADAGECasey Greene · Network Biology 2016
• 46mScalable Genomics with ADAMAndy Petrella, Xavier Tordoir · Scala.IO 2014
• 30mLeveraging Molecular Data for Drug DiscoveryHaiyan Huang · Computational Challenges in Very Large-Scale 'Omics'
• 28mNanopore Basecalling for Directed EvolutionIan Holmes · Computational Challenges in Very Large-Scale 'Omics'
• 01h 29mFundamental Algorithms in Deep Sequencing III: Genomic Compression: Storage, Transmission, and ...Noah M. Daniels · Algorithmic Challenges in Genomics Boot Camp
• 31mHacking the Human Genome Using Clojure and Similarity SearchArnoldo Jose Muller-Molina · Clojure/conj 2011