Refinement Contrastive Learning of Cell-Gene Associations for Unsupervised Cell Type Identification

arXiv — cs.LGFriday, December 12, 2025 at 5:00:00 AM
  • A new framework called Refinement Contrastive Learning (scRCL) has been proposed to enhance unsupervised cell type identification by incorporating cell-gene associations. This approach aims to improve the representation of cellular structures by utilizing contrastive distribution alignment components and a refinement module that captures gene-correlation structures.
  • The development of scRCL is significant as it addresses limitations in existing clustering methods that often overlook the importance of cell-gene interactions, thereby enhancing the accuracy of identifying closely related cell types in single-cell omics studies.
  • This advancement aligns with ongoing efforts in the field of artificial intelligence to improve biological data analysis, particularly in RNA-seq and spatial transcriptomics. The integration of gene associations into cell type identification reflects a broader trend towards more nuanced and informative models that can better capture the complexities of cellular environments.
— via World Pulse Now AI Editorial System

Was this article worth reading? Share it

Recommended apps based on your readingExplore all apps
Continue Readings
HEIST: A Graph Foundation Model for Spatial Transcriptomics and Proteomics Data
PositiveArtificial Intelligence
A new framework named HEIST has been introduced to enhance the analysis of spatial transcriptomics and proteomics data, addressing the limitations of existing models that overlook spatial information and complex cellular programs. This model aims to provide insights into cellular heterogeneity and gene expression at the single-cell level by incorporating spatial coordinates and intra-cellular counts.
STARK denoises spatial transcriptomics images via adaptive regularization
PositiveArtificial Intelligence
A novel method called STARK has been developed to denoise spatial transcriptomics images, enhancing the identification of cell types at ultra-low sequencing depths. This technique employs adaptive regularization and kernel ridge regression, updating a spatial graph iteratively to improve image quality and gene expression interpolation.

Ready to build your own newsroom?

Subscribe to unlock a personalised feed, podcasts, newsletters, and notifications tailored to the topics you actually care about