MLRR-ATV: A Robust Manifold Nonnegative Low-Rank Representation With Adaptive Total-Variation Regularization for...

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Single-cell RNA sequencing (scRNA-seq) enables gene expression analysis at the individual cell level.
• scRNA-seq data is characterized by high dimensionality, sparsity, and significant noise due to technological limitations.
• Clustering is a fundamental technique for analyzing scRNA-seq data to identify cell populations.

Purpose of the Study:

• To develop a novel and robust method for clustering scRNA-seq data.
• To address the challenges of noise, high dimensionality, and sparsity in scRNA-seq datasets.
• To improve the accuracy and reliability of cell type identification through advanced clustering.

Main Methods:

• Introduction of a novel Robust Manifold Nonnegative Low-Rank Representation with Adaptive Total-Variation Regularization (MLRR-ATV) method.
• Integration of Adaptive Total-Variation (ATV) regularization within a Low-Rank Representation (LRR) framework to mitigate noise via gradient learning.
• Incorporation of Euclidean distance and cosine similarity to capture both linear and nonlinear manifold structures within the data.
• Utilization of the Alternating Direction Method of Multipliers (ADMM) for optimizing the non-convex MLRR-ATV model.

Main Results:

• The MLRR-ATV model demonstrated superior performance compared to nine state-of-the-art methods across eight real-world scRNA-seq datasets.
• The method effectively reduced the influence of noise, preserving crucial biological information within the datasets.
• Accurate identification of cell populations was achieved, highlighting the model's effectiveness in scRNA-seq data analysis.

Conclusions:

• MLRR-ATV offers a significant advancement in single-cell RNA sequencing data clustering.
• The proposed method provides a robust and accurate solution for analyzing noisy, high-dimensional, and sparse single-cell data.
• MLRR-ATV enhances the ability to explore gene expression and identify cell types, contributing to a deeper understanding of cellular heterogeneity.