This paper introduces scDHA, a single-cell Decomposition using Hierarchical Autoencoder framework, designed to efficiently analyze scRNA-seq data by extracting representative information from each cell while mitigating noise. scDHA comprises two modules: a non-negative kernel autoencoder for removing insignificant genes and a stacked Bayesian autoencoder for projecting data onto a lower-dimensional space. The framework outperforms state-of-the-art techniques in cell segregation, visualization, classification, and pseudo-time inference.
Publisher
Nature Communications
Published On
Feb 15, 2021
Authors
Duc Tran, Hung Nguyen, Bang Tran, Carlo La Vecchia, Hung N. Luu, Tin Nguyen
Tags
single-cell analysis
scRNA-seq
noise reduction
autoencoder
data visualization
classification
pseudo-time inference
Related Publications
Explore these studies to deepen your understanding of the subject.
