EUVIMEDEUVIMED entdecken
Open Access

Matrix sketching framework for linear mixed models in association studies

Myson Burch iD, Aritra Bose iD, Gregory Dexter, Laxmi Parida iD, Petros Drineas iD

Genome Research · 2024

DOI: 10.1101/gr.279230.124 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Genome Research, 2024. Verfasst von Myson Burch und 4 weiteren Autor:innen. Beteiligte Einrichtung: Computational Genomics, IBM T.J. Watson Research Center, Yorktown Heights, New York 10598, USA. Hauptthemen laut MEDLINE: Genome-Wide Association Study. Die Arbeit nennt 26 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: Division of Information and Intelligent Systems, National Science Foundation, IBM Faculty Award.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Linear mixed models (LMMs) have been widely used in genome-wide association studies to control for population stratification and cryptic relatedness. However, estimating LMM parameters is computationally expensive, necessitating large-scale matrix operations to build the genetic relationship matrix (GRM). Over the past 25 years, Randomized Linear Algebra has provided alternative approaches to such matrix operations by leveraging matrix sketching , which often results in provably accurate fast and efficient approximations. We leverage matrix sketching to develop a fast and efficient LMM method called Ma trix- Sk etching LMM (MaSk-LMM) by sketching the genotype matrix to reduce its dimensions and speed up computations. Our framework comes with both theoretical guarantees and a strong empirical performance compared to the current state-of-the-art for simulated traits and complex diseases.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Genome-Wide Association StudyHumansLinear ModelsModels, GeneticGenotypeAlgorithmsComputer Simulation

Autor:innen und Institutionen

  • Computational Genomics, IBM T.J. Watson Research Center, Yorktown Heights, New York 10598, USA.
  • Computer Science Department, Purdue University, West Lafayette, Indiana 47907, USA.
  • Computer Science Department, Purdue University, West Lafayette, Indiana 47907, USA pdrineas@purdue.edu.

ORCID-Profile: Myson Burch, Aritra Bose, Laxmi Parida, Petros Drineas

Förderung

  • Division of Information and Intelligent Systems
  • National Science Foundation
  • IBM Faculty Award
  • NSF Graduate Research Fellowship Program
  • GRFP
  • IBM Research

Zitieren

APA

Burch, M., Bose, A., Dexter, G., Parida, L., & Drineas, P. (2024). Matrix sketching framework for linear mixed models in association studies. Genome Research. https://doi.org/10.1101/gr.279230.124

Vancouver

Burch M, Bose A, Dexter G, Parida L, Drineas P. Matrix sketching framework for linear mixed models in association studies. Genome Research. 2024. doi:10.1101/gr.279230.124

BibTeX
@article{Burch2024,
  author = {Burch, Myson and Bose, Aritra and Dexter, Gregory and Parida, Laxmi and Drineas, Petros},
  title = {Matrix sketching framework for linear mixed models in association studies},
  journal = {Genome Research},
  year = {2024},
  doi = {10.1101/gr.279230.124}
}
Kostenlos starten →