Related Experiment Video
Updated: Apr 25, 2026
Author Spotlight: Impact of Intergenic Interactions on Disease-Identifying Dark Biomarkers
Published on: March 1, 2024
Locally linear representation Fisher criterion based tumor gene expressive data classification
Bo Li1, Bei-Bei Tian2, Xiao-Long Zhang2
1School of Computer Science and Technology, Wuhan University of Science and Technology, Wuhan, Hubei 430065, China; Hubei Province Key Laboratory of Intelligent Information Processing and Real-time Industrial System, Wuhan, Hubei 430065, China; Department of Electrical and Computer Engineering, Ryerson University, Toronto, Ontario, Canada M5B 2K3.
This study introduces a new method, locally linear representation Fisher criterion (LLRFC), for dimensionality reduction in tumor gene expression data. LLRFC effectively extracts features, improving the identification of tumor subtypes.
Area of Science:
- Bioinformatics
- Computational Biology
- Genomics
Background:
- Tumor gene expression data present high dimensionality challenges due to numerous genes and limited samples.
- Effective dimensionality reduction is crucial for accurate tumor subtype identification and analysis.
Purpose of the Study:
- To introduce and validate a novel discriminant manifold learning method for tumor gene expression data.
- To enhance feature extraction for improved tumor classification.
Main Methods:
- Locally Linear Representation Fisher Criterion (LLRFC) was developed for feature extraction.
- LLRFC constructs inter-class and intra-class graphs using k-nearest neighbors with distinct or same class labels.
- Locally least linear reconstruction and Fisher criterion optimize weights and identify a low-dimensional subspace.
Main Results:
- LLRFC demonstrated efficiency in experiments on benchmark tumor gene expression datasets.
- The method effectively maximizes inter-class reconstruction errors and minimizes intra-class reconstruction errors.
- Validation against related algorithms confirmed the efficacy of LLRFC.
Conclusions:
- LLRFC is an efficient and effective method for dimensionality reduction in tumor gene expression data.
- The proposed approach improves feature extraction for better tumor subtype identification.
- This method offers a valuable tool for genomic data analysis in cancer research.
