Machine learning-based method for associating gene expression with cell microscopic phenotype

By combining large-scale models in the visual domain with spatial transcriptome data for single-cell segmentation and machine learning analysis, the problem of high data analysis difficulty in phenotypic identification is solved. This enables efficient association between microscopic phenotypes and gene expression, improves cell segmentation accuracy, and enhances the understanding of the impact of gene expression on cell morphology and function.

WO2026102819A1PCT designated stage Publication Date: 2026-05-21THE INST OF BIOTECHNOLOGY OF THE CHINESE ACAD OF AGRI SCI
View PDF 4 Cites 0 Cited by

Patent Information

Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
THE INST OF BIOTECHNOLOGY OF THE CHINESE ACAD OF AGRI SCI
Filing Date
2024-11-29
Publication Date
2026-05-21

AI Technical Summary

Technical Problem

Existing phenotypic identification methods present significant challenges in data analysis, making it difficult to effectively correlate single-cell microscopic phenotypes with gene expression.

Method used

By fine-tuning a large-scale vision model and combining it with spatial transcriptome data for single-cell segmentation, cell imaging data and spatial transcriptome data are obtained. Then, a machine learning model is used for correlation analysis to determine cell mask and gene expression, thus realizing the correlation between microscopic phenotype and gene expression.

Benefits of technology

It improves the accuracy and efficiency of cell segmentation, accurately matches gene expression information with cell location, reveals how gene expression affects cell morphology and function, and solves the technical problem of difficult data analysis.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN2024135706_21052026_PF_FP_ABST
    Figure CN2024135706_21052026_PF_FP_ABST
Patent Text Reader

Abstract

Provided in the present invention is a machine learning-based method for associating gene expression with cell microscopic phenotype, the method comprising: acquiring cell imaging data and spatial transcriptome data, the cell imaging data comprising a cell nucleus image of a target cell; by means of fine-tuning a vision domain large model and on the basis of the cell nucleus image and the spatial transcriptome data, performing single-cell segmentation to obtain a cell segmentation image of the target cell; on the basis of the cell segmentation image, determining a cell mask of the target cell; quantifying the cell mask to obtain the microscopic phenotype of the target cell; on the basis of the cell mask of the target cell and the spatial transcriptome data, determining the gene expression of the target cell; and, by means of a preset machine learning model and on the basis of the microscopic phenotype of the target cell and the gene expression of the target cell, performing association analysis to obtain a result of association between the microscopic phenotype of the target cell and the gene expression thereof. The present invention can associate the microscopic phenotype of a single cell with the gene expression thereof.
Need to check novelty before this filing date? Find Prior Art