Predicting the Deleteriousness of Genomic Variants – Big and Small

on YouTube ↗27m

" summary

CADD scores short variants by contrasting fixed human‑derived alleles with simulated de novo changes and integrates many annotations on a single continuous scale, and sequence‑based DNN splice predictors like SpliceAI and MMSplice are added to boost splice‑variant ranking while failing to capture concurrent nonsense/missense effects. The same contrastive framework extends to CADD-SV, which summarizes genome annotations across SV spans, outperforms SVScore/AnnotSV on pathogenicity and frequency, and currently omits inversions/translocations and inserted‑sequence modeling.

AI-written summary · about

" chapters (18)

1. 0:00Intro
2. 0:22Genome-wide predictors of variant effects
3. 1:15Precision medicine: Sequence variant interpretation
4. 1:41Variant filtering hierarchies
5. 2:08No annotation vs. annotation overload
6. 2:23Data integration with a supervised machine learning approach
7. 2:51Clinical variant sources
8. 3:41Clin Var lacks exon distal variants
9. 4:45Representation of non-coding variants does not improve
10. 6:42Purifying selection creates annotation dependent depletion
11. 7:15CADD for variant prioritization
12. 8:44Integration of sequence derived scores
13. 9:22Integrating Splice Deep Learning models in v1.6 ("CADD-Splice")
14. 11:15Importance of regulatory sequences
15. 12:39Performance of computational predictors on MPRA data
16. 14:45Goal: Extended sequence based models of regulatory effects
17. 16:32Structural variants (SVS) & disease Insertions, deletions, inversions and translocations, Copy Number Variants
18. 20:56Limitations and outlook

" related talks

• 05mDecoding the Human Genome with Deep LearningOlga Troyanskaya · O'Reilly Artificial Intelligence Conference 2019 - New York, New York
• 01h 31mIntroduction to Cancer Bioinformatics I: Inferring Genomic Variation from Tumor Sequencing DataBen Raphael, Niko Beerenwinkel · Algorithmic Challenges in Genomics Boot Camp
• 38mInterpreting Deep Learning Models Of Functional Genomics Data To Decode Regulatory Sequence...Anshul Kundaje · Interpretable Machine Learning in Natural and Social Sciences (2022)
• 30mLeveraging Molecular Data for Drug DiscoveryHaiyan Huang · Computational Challenges in Very Large-Scale 'Omics'
• 39mIdentifying Disease Mechanisms Using Gene Networks as PriorsAnna Goldberg · Network Biology 2016
• 12mHow data science can make your genome actionableRiley Ennis · Collision 2017
• 20mDiagnostic Medical application of MLCarrie Weinberg · MLconf NYC 2019
• 12mEnhancing Code Vulnerability Detection via Vulnerability-Preserving Data AugmentationYang Liu, Xiaofei Xie, Shangqing Liu, Wei Ma, Jian Wang, Ruitao Feng · LCTES 2024