EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Accurately modeling biased random walks on weighted networks using node2vec+

Renming Liu iD, Matthew Hirn iD, Arjun Krishnan iD

Bioinformatics · 01.01.2023 · 39 (1)

DOI: 10.1093/bioinformatics/btad047 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Bioinformatics, online erschienen am 23.01.2023. Verfasst von Renming Liu und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Computational Mathematics, Science & Engineering, Michigan State University, East Lansing, MI 48824, USA. Hauptthemen laut MEDLINE: Neural Networks, Computer, Machine Learning. Die Arbeit wurde bisher 34-mal zitiert (Crossref) und nennt 47 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: NIGMS NIH HHS, NSF, NIH.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Motivation Accurately representing biological networks in a low-dimensional space, also known as network embedding, is a critical step in network-based machine learning and is carried out widely using node2vec, an unsupervised method based on biased random walks. However, while many networks, including functional gene interaction networks, are dense, weighted graphs, node2vec is fundamentally limited in its ability to use edge weights during the biased random walk generation process, thus under-using all the information in the network. Results Here, we present node2vec+, a natural extension of node2vec that accounts for edge weights when calculating walk biases and reduces to node2vec in the cases of unweighted graphs or unbiased walks. Using two synthetic datasets, we empirically show that node2vec+ is more robust to additive noise than node2vec in weighted graphs. Then, using genome-scale functional gene networks to solve a wide range of gene function and disease prediction tasks, we demonstrate the superior performance of node2vec+ over node2vec in the case of weighted graphs. Notably, due to the limited amount of training data in the gene classification tasks, graph neural networks such as GCN and GraphSAGE are outperformed by both node2vec and node2vec+. Availability and implementation The data and code are available on GitHub at https://github.com/krishnanlab/node2vecplus_benchmarks. All additional data underlying this article are available on Zenodo at https://doi.org/10.5281/zenodo.7007164. Supplementary information Supplementary data are available at Bioinformatics online.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Neural Networks, ComputerMachine LearningGene Regulatory NetworksPhenotypeEpistasis, Genetic

Autor:innen und Institutionen

  • Department of Computational Mathematics, Science & Engineering, Michigan State University, East Lansing, MI 48824, USA.
  • Department of Mathematics, Michigan State University, East Lansing, MI 48824, USA.
  • Center for Quantum Computing, Science & Engineering, Michigan State University, East Lansing, MI 48824, USA.
  • Department of Biomedical Informatics, University of Colorado Anschutz Medical Campus, Aurora, CO 80045, USA.

ORCID-Profile: Renming Liu, Matthew Hirn, Arjun Krishnan

Förderung

  • NIGMS NIH HHS
  • NSF
  • NIH

Zitieren

APA

Liu, R., Hirn, M., & Krishnan, A. (2023). Accurately modeling biased random walks on weighted networks using node2vec+. Bioinformatics, 39(1). https://doi.org/10.1093/bioinformatics/btad047

Vancouver

Liu R, Hirn M, Krishnan A. Accurately modeling biased random walks on weighted networks using node2vec+. Bioinformatics. 2023;39(1). doi:10.1093/bioinformatics/btad047

BibTeX
@article{Liu2023,
  author = {Liu, Renming and Hirn, Matthew and Krishnan, Arjun},
  title = {Accurately modeling biased random walks on weighted networks using node2vec+},
  journal = {Bioinformatics},
  year = {2023},
  volume = {39},
  number = {1},
  doi = {10.1093/bioinformatics/btad047}
}
Kostenlos starten →