EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

rPLS-LDA: A Robust Partial Least Squares Linear Discriminant Analysis for Transcriptomics and Metabolomics Data Analysis

Md. Rabiul Auwul, Md. Shahjaman

Letters in Applied NanoBioScience · 30.03.2026 · 15 (1)

DOI: 10.33263/lianbs151.034 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Letters in Applied NanoBioScience, online erschienen am 30.03.2026. Verfasst von Md. Rabiul Auwul und 1 weiteren Autor:innen. Die Arbeit nennt 60 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Classification is a supervised learning approach for allocating unlabeled dichotomous patient samples into labeled dichotomous patient samples. Conventional classification methods are clinically based and therefore have limited applications, with lower prediction accuracy. Partial Least Squares Linear Discriminant Analysis (PLS-LDA) is one of the most popular classifiers for predicting unlabeled dichotomous patient samples. Nevertheless, this method yields misleading results when an outlying data vector is present under classical maximum likelihood estimators. In this paper, we developed a robust PLS-LDA approach using the minimum beta divergence method. The efficiency of the proposed approach was evaluated through comparisons with existing classification methods using both simulated and real datasets (transcriptomics and metabolomics datasets). Results from 10-fold cross-validation showed that the proposed method significantly outperforms other methods across various performance indices at three outlier rates (10%, 20%, and 30%). Whereas, in the absence of outliers, it keeps almost equal performance with the traditional PLS-LDA method. The protein-protein interaction (PPI), Gene Ontology (GO), and KEGG analyses of the 10 genes identified in prostate cancer data by the proposed procedure demonstrated the significance of the proposed methods. The computational tool has been implemented in an R package, which is publicly available from https://github.com/MdRabiulAuwul/rPLS-LDA.

Zitieren

APA

Auwul, M. R., & Shahjaman, M. (2026). rPLS-LDA: A Robust Partial Least Squares Linear Discriminant Analysis for Transcriptomics and Metabolomics Data Analysis. Letters in Applied NanoBioScience, 15(1). https://doi.org/10.33263/lianbs151.034

Vancouver

Auwul MR, Shahjaman M. rPLS-LDA: A Robust Partial Least Squares Linear Discriminant Analysis for Transcriptomics and Metabolomics Data Analysis. Letters in Applied NanoBioScience. 2026;15(1). doi:10.33263/lianbs151.034

BibTeX
@article{Auwul2026,
  author = {Auwul, Md. Rabiul and Shahjaman, Md.},
  title = {rPLS-LDA: A Robust Partial Least Squares Linear Discriminant Analysis for Transcriptomics and Metabolomics Data Analysis},
  journal = {Letters in Applied NanoBioScience},
  year = {2026},
  volume = {15},
  number = {1},
  doi = {10.33263/lianbs151.034}
}
Kostenlos starten →