Quantitative Modeling of Stemness in Single-Cell RNA Sequencing Data: A Nonlinear One-Class Support Vector Machine...

Area of Science:

• Computational biology
• Genomics
• Cancer research

Background:

• Intratumoral heterogeneity and cancer stem cells pose significant challenges in cancer therapy.
• Quantifying cellular stemness is vital for understanding cancer evolution and developing targeted therapies.
• Accurate measurement of stemness is difficult using current molecular data.

Purpose of the Study:

• To propose a novel method for quantifying cell stemness using single-cell RNA sequencing (scRNA-seq) data.
• To introduce the one-class Hadamard kernel support vector machine (OCHSVM) for stemness definition.
• To evaluate the OCHSVM method's performance against existing stemness identification techniques.

Main Methods:

• Development of the one-class Hadamard kernel support vector machine (OCHSVM) algorithm.
• Application of OCHSVM to analyze scRNA-seq data from various cell types (embryo, iPSC, tumor).
• Comparative analysis of OCHSVM against CytoTRACE, one-class logistic regression, and other SVM methods.

Main Results:

• The OCHSVM method demonstrated suitability for stemness identification in scRNA-seq data.
• OCHSVM showed improved performance compared to state-of-the-art methods in computational assessments.
• The study validated OCHSVM's effectiveness across diverse biological datasets.

Conclusions:

• The proposed OCHSVM method offers a robust approach for quantifying cell stemness.
• OCHSVM provides a valuable tool for analyzing cancer stemness from scRNA-seq data.
• This advancement aids in understanding cancer dynamics and developing targeted therapies.