Segmentation of NKX2.5 Signal in Human Pluripotent Stem Cell-Derived Cardiomyocytes

Siem Jongsma, Verena Schwach, Simone A.Ten Den, Robert Passier, Fons J. Verbeek, Lu Cao*

*Corresponding author for this work

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

Abstract

Human pluripotent stem cell-derived Cardiomyocytes (hPSC-CMs) become increasingly popular in recent years for disease modeling and drug screening. NKX2.5 gene is a key transcription factor that regulates cardiomyocyte differentiation. A human embryonic stem cell (hESC) reporter line with NKX2.5 in GFP signal allows us to monitor the specificity and efficiency of human cardiac differentiation. We intend to develop an automatic analysis pipeline for the NKX2.5 signal. However, the NKX2.5 signal captured from fluorescence microscopy is highly heterogeneous. It is not possible to be properly segmented using traditional thresholding methods. Therefore, in this paper, one machine learning method: enhanced Fuzzy C-Means clustering (EnFCM) and two deep learning models: U-Net and DeepLabV3+, are evaluated on the segmentation performance. Parameters have been tuned for each method so as to reach to the optimal segmentation performance. The results show that EnFCM reaches the performance of 0.85. U-Net and DeepLabV3+ have a superior performance. Their performances are 0.86 and 0.89 respectively.

Original languageEnglish
Title of host publicationData Science and Artificial Intelligence - 1st International Conference, DSAI 2023, Proceedings
EditorsChutiporn Anutariya, Marcello M. Bonsangue
PublisherSpringer
Pages170-184
Number of pages15
ISBN (Electronic)978-981-99-7969-1
ISBN (Print)978-981-99-7968-4
DOIs
Publication statusPublished - 2023
Event1st International Conference on Data Science and Artificial Intelligence, DSAI 2023 - Bangkok, Thailand
Duration: 27 Nov 202329 Nov 2023
Conference number: 1

Publication series

NameCommunications in Computer and Information Science
Volume1942 CCIS
ISSN (Print)1865-0929
ISSN (Electronic)1865-0937

Conference

Conference1st International Conference on Data Science and Artificial Intelligence, DSAI 2023
Abbreviated titleDSAI 2023
Country/TerritoryThailand
CityBangkok
Period27/11/2329/11/23

Keywords

  • deep learning
  • machine learning
  • pluripotent stem cell derived cardiomyocyte
  • segmentation
  • 2024 OA procedure

Fingerprint

Dive into the research topics of 'Segmentation of NKX2.5 Signal in Human Pluripotent Stem Cell-Derived Cardiomyocytes'. Together they form a unique fingerprint.

Cite this