GenVarLoader: An accelerated dataloader for applying deep learning to personalized genomics

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 17, 2025

Abstract Deep learning sequence models trained on personalized genomics can improve variant effect prediction, however, applications of these are limited by computational requirements for storing and reading large datasets. We address this with GenVarLoader, which stores genomic data in new memory-mapped formats optimal locality to achieve ∼1,000x faster throughput ∼2,000x better compression compared existing alternatives.

Language: Английский

---

Deep Genomic Models of Allele-Specific Measurements

Published: Jan. 1, 2025

Language: Английский

---

Buffering and non-monotonic behavior of gene dosage response curves for human complex traits

medRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 11, 2024

A bstract The genome-wide burdens of deletions, loss-of-function mutations, and duplications correlate with many traits. Curiously, for most these traits, variants that decrease expression have the same average direction effect as increase expression. This seemingly contradicts intuition that, at individual genes, reducing should opposite on a phenotype increasing To understand this paradox, we introduce concept called gene dosage response curve (GDRC) relates changes in to expected phenotype. We show GDRCs are systematically biased one trait relative other and, surprisingly, 40% non-monotone, large increases decreases affecting direction. develop simple theoretical model explains bias Our results broad implications complex drug discovery, statistical genetics.

Language: Английский

---

Split-pool barcoding serves up an epigenomic smorgasbord

Nature Genetics, Journal Year: 2024, Volume and Issue: 56(12), P. 2596 - 2597

Published: Nov. 25, 2024

Language: Английский

---