Skip to main navigation Skip to search Skip to main content

ScSSC: Semi-supervised Single Cell Clustering Based on 2D Embedding

  • Harbin Institute of Technology
  • School of Computer Science and Technology, Harbin Institute of Technology

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In recent years, with the development of single-cell RNA sequencing (scRNA-seq) technology, more and more scRNA-seq data has been generated. Corresponding analysis methods such as clustering analysis are also proposed, which effectively distinguish the cell types and reveal the cell diversity. However, due to more than ten thousand genes for normal species, the dimension of scRNA-seq data is very high. Meanwhile, there exist many zero counts in scRNA-seq data. They all increase the difficulty of clustering analysis of scRNA-seq data. This paper proposes ScSSC, a semi-supervised clustering method based on 2D embedding. ScSSC uses the autoencoder for pre-training to construct the network and applies the community discovery algorithm to label cells. Then a semi-supervised network is used to clustering the data after training. The clustering results of three public data sets show that ScSSC has better performance than other clustering methods.

Original languageEnglish
Title of host publicationIntelligent Computing Theories and Application - 17th International Conference, ICIC 2021, Proceedings
EditorsDe-Shuang Huang, Kang-Hyun Jo, Jianqiang Li, Valeriya Gribova, Vitoantonio Bevilacqua
PublisherSpringer Science and Business Media Deutschland GmbH
Pages478-489
Number of pages12
ISBN (Print)9783030845315
DOIs
StatePublished - 2021
Externally publishedYes
Event17th International Conference on Intelligent Computing, ICIC 2021 - Shenzhen, China
Duration: 12 Aug 202115 Aug 2021

Publication series

NameLecture Notes in Computer Science
Volume12838 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference17th International Conference on Intelligent Computing, ICIC 2021
Country/TerritoryChina
CityShenzhen
Period12/08/2115/08/21

Keywords

  • 2D embedding
  • Autoencoder
  • Community discovery
  • Semi-supervised learning
  • Single-cell clustering

Fingerprint

Dive into the research topics of 'ScSSC: Semi-supervised Single Cell Clustering Based on 2D Embedding'. Together they form a unique fingerprint.

Cite this