Answers are grounded in this paper and cite their sources.
Computational study demonstrates improved biomarker selection and survival prediction across cancer cohorts, indicating enhanced multi-omics integration capabilities.
Introduction: With the advancement of high-throughput studies, an increasing wealth of high-dimensional multi-omics data is being collected from the same patient cohort. However, leveraging this multi-omics data to predict survival outcomes poses a significant challenge due to its complex structure. Methods: In this article, we present a novel approach, the Adaptive Sparse Multi-Block Partial Least Squares (asmbPLS) Regression model, which introduces a dynamic assignment of penalty factors to distinct blocks within various PLS components, facilitating effective feature selection and prediction. Results: We compared the proposed method with several state-of-the-art algorithms encompassing prediction performance, feature selection and computation efficiency. We conducted comprehensive evaluations using both simulated data with various scenarios and a real dataset from the melanoma patients to validate the effectiveness and efficiency of the asmbPLS method. Additionally, we applied the lung squamous cell carcinoma (LUSC) dataset from The Cancer Genome Atlas (TCGA) to further assess the feature selection capability of asmbPLS. Discussion: The inherent nature of asmbPLS imparts it with higher sensitivity in feature selection compared to other methods. Furthermore, an R package called asmbPLS implementing this method is made publicly available.
No takes yet. Share an insight, caveat, or question.
Zhang et al. (2024) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: