EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Multiple instance fine-mapping: Predicting causal regulatory variants with a deep sequence model

Alexander Rakowski iD, Christoph Lippert iD

PLOS Genetics · 29.06.2026 · 22 (6) · e1012208

DOI: 10.1371/journal.pgen.1012208 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS Genetics, online erschienen am 29.06.2026. Verfasst von Alexander Rakowski und 1 weiteren Autor:innen. Beteiligte Einrichtung: Digital Health Machine Learning, Hasso Plattner Institute for Digital Engineering, Potsdam, Germany. Hauptthemen laut MEDLINE: Genome-Wide Association Study, Chromosome Mapping. Die Arbeit nennt 62 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Horizon 2020.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Identifying causal genetic variants in a computational manner remains an open problem. Training end-to-end prediction models is not possible without large ground-truth datasets, while results of genome-wide association studies (GWAS) are entangled by linkage disequilibrium (LD), and gene expression datasets do not contain genetic variation at individual-level. Here, we propose Multiple Instance Fine-mapping (MIFM) – a multiple instance learning (MIL) objective to overcome the lack of strong labels by grouping putatively causal variants together based on their LD scores. Using MIFM, we trained a deep classifier on a dataset aggregating over 13,000 GWAS to predict causal variants based on their underlying DNA sequences. We validated variants prioritized by MIFM by constructing polygenic risk scores which transferred better to different target ancestries. Furthermore, we demonstrated how MIFM can be used to disentangle effect sizes of highly-correlated variants to better fine-map GWAS results.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Genome-Wide Association StudyChromosome MappingLinkage DisequilibriumHumansPolymorphism, Single NucleotideGenetic Risk ScoreGenetic VariationMultifactorial InheritanceModels, Genetic

Autor:innen und Institutionen

  • Digital Health Machine Learning, Hasso Plattner Institute for Digital Engineering, Potsdam, Germany.
  • Hasso Plattner Institute for Digital Health, Icahn School of Medicine at Mount Sinai, New York, New York, United States of America.

ORCID-Profile: Alexander Rakowski, Christoph Lippert

Förderung

  • Horizon 2020

Zitieren

APA

Rakowski, A., & Lippert, C. (2026). Multiple instance fine-mapping: Predicting causal regulatory variants with a deep sequence model. PLOS Genetics, 22(6), e1012208. https://doi.org/10.1371/journal.pgen.1012208

Vancouver

Rakowski A, Lippert C. Multiple instance fine-mapping: Predicting causal regulatory variants with a deep sequence model. PLOS Genetics. 2026;22(6):e1012208. doi:10.1371/journal.pgen.1012208

BibTeX
@article{Rakowski2026,
  author = {Rakowski, Alexander and Lippert, Christoph},
  title = {Multiple instance fine-mapping: Predicting causal regulatory variants with a deep sequence model},
  journal = {PLOS Genetics},
  year = {2026},
  volume = {22},
  number = {6},
  pages = {e1012208},
  doi = {10.1371/journal.pgen.1012208}
}
Kostenlos starten →