IP Library › Granted Patent US 12,525,317
Granted Patent B2
US 12,525,317 · App. 18/555,419 · Granted Jan 13, 2026

Apparatus and method for predicting cell type enrichment from tissue images using spatially resolved gene expression data

Inventors: Hongyoon Choi (Seoul, KR); Young Tae Kim (Seoul, KR); Kwon Joong Na (Seoul, KR)
Assignee: PORTRAI INC.
G16B20/00G06T7/0012G06V10/774G06V20/695G06V20/698G16B25/10G16B40/20G06T2207/20081G06T2207/30024
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,525,317
App. No.
18/555,419
Granted
Jan 13, 2026
Kind
B2
Abstract

An apparatus for predicting cell type enrichment from tissue images based on spatially resolved gene expression information comprises: a communication module receiving tissue images of a test object; a memory storing a program for predicting cell type enrichment information from the tissue images; and a processor executing the program. The program predicts the cell type enrichment information by inputting the tissue images to a cell type enrichment prediction model which is trained using training data that consists of spatially resolved transcriptome information and tissue images spatially aligned with the spatially resolved transcriptome information. The spatially resolved transcriptome information includes: transcriptome data including spatial information; and tissue image data sharing the spatial information. The spatial information refers to position information about multiple spots arranged in a two-dimensional plane in the tissue image data, and includes the coordinates of each of the spots.

Claims (28)

1 . An apparatus for predicting cell type enrichment from tissue images based on spatially resolved gene expression information, comprising:

a communication module receiving tissue images of a specimen;

a memory storing a program for predicting cell type enrichment information from the tissue images; and

a processor executing the program,

wherein the program predicts the cell type enrichment information by inputting the tissue images to a cell type enrichment prediction model which is trained using training data that consists of spatially resolved transcriptome information and tissue images spatially aligned with the spatially resolved transcriptome information,

wherein the spatially resolved transcriptome information includes transcriptome data having spatial information and tissue image data sharing the spatial information,

wherein the spatial information means position information about multiple spots arranged in a two-dimensional plane in the tissue image data and includes coordinates of each of the spots.

2 . The apparatus for predicting cell type enrichment from tissue images according to claim 1 , wherein the cell type enrichment prediction model is constructed based on training data in which the spatially resolved transcriptome information for each of previously collected human or animal tissues is matched to the cell type enrichment information with respect to the transcriptome data classified according to the coordinates of each of the spots.

3 . The apparatus for predicting cell type enrichment from tissue images according to claim 2 , wherein the cell type enrichment prediction model includes: an image dividing part dividing the tissue image data into patch tissue images of a preset size; a molecule marker model part outputting the cell distribution information for each type of cell group labeled through the transcriptome data, and a prediction part extracting the cell type enrichment information labeled with the transcriptome data based on the coordinates of a central spot among the multiple spots present in the patch tissue images.

4 . The apparatus for predicting cell type enrichment from tissue images according to claim 3 , wherein the image dividing part of the cell type enrichment prediction model is constructed to perform a process of matching the transcriptome data with the tissue image data based on the coordinates of the spots, a process of placing a rectangular box of a preset size in the tissue image data including the multiple spots, and a process of extracting the tissue image data into at least one patch tissue image such that the coordinates of the central spot among the multiple spots become central coordinates of the rectangular box.

5 . The apparatus for predicting cell type enrichment from tissue images according to claim 3 , wherein the molecule marker model part of the cell type enrichment prediction model is constructed based on training data in which cell distribution information for each type of cell group present in existing tissue images published on the Internet is matched to transcriptome data present in each single-cell type.

6 . The apparatus for predicting cell type enrichment from tissue images according to claim 3 , wherein the program includes information about the types of cell groups predicted from the tissue images and a heatmap tissue image indicating distribution information of cells in each type of cell group, as the cell type enrichment information.

7 . A method for predicting cell type enrichment from tissue images based on spatially resolved gene expression information using an apparatus for predicting cell type enrichment from tissue images, the method comprising:

receiving tissue images of a specimen; and

predicting cell type enrichment information by inputting the tissue images to a cell type enrichment prediction model which is trained using training data that consists of spatially resolved transcriptome information and tissue images spatially aligned with the spatially resolved transcriptome information,

wherein the spatially resolved transcriptome information includes transcriptome data having spatial information and tissue image data sharing the spatial information, and

wherein the spatial information means position information about multiple spots arranged in a two-dimensional plane in the tissue image data and includes coordinates of each of the spots.

8 . The method for predicting cell type enrichment from tissue images according to claim 7 , wherein the cell type enrichment prediction model is constructed based on training data in which the spatially resolved transcriptome information for each of previously collected human or animal tissues is matched to the cell type enrichment information with respect to the transcriptome data classified according to the coordinates of each of the spots.

9 . The method for predicting cell type enrichment from tissue images according to claim 8 , wherein the cell type enrichment prediction model includes: an image dividing part dividing the tissue image data into patch tissue images of a preset size; a molecule marker model part outputting the cell distribution information for each type of cell group labeled through the transcriptome data, and a prediction part extracting the cell type enrichment information labeled with the transcriptome data based on the coordinates of a central spot among the multiple spots present in the patch tissue images.

10 . The method for predicting cell type enrichment from tissue images according to claim 9 , wherein the molecule marker model part is constructed based on training data in which cell distribution information for each type of cell group present in existing tissue images published on the Internet is matched to transcriptome data present in each single-cell type.

11 . A non-transitory computer-readable recording medium in which a computer program for executing the method for predicting cell type enrichment from tissue images according to claim 10 is recorded.

12 . The method for predicting cell type enrichment from tissue images according to claim 9 , wherein, in the step of predicting cell type enrichment information, the cell type enrichment information includes information about the types of cell groups predicted from the tissue images and a heatmap tissue image indicating distribution information of cells in each type of cell group.

13 . A non-transitory computer-readable recording medium in which a computer program for executing the method for predicting cell type enrichment from tissue images according to claim 12 is recorded.

14 . A non-transitory computer-readable recording medium in which a computer program for executing the method for predicting cell type enrichment from tissue images according to claim 9 is recorded.

15 . The method for predicting cell type enrichment from tissue images according to claim 8 , wherein the image dividing part of the cell type enrichment prediction model is constructed to perform a process of matching the transcriptome data with the tissue image data based on the coordinates of the spots, a process of placing a rectangular box of a preset size in the tissue image data including the multiple spots, and a process of extracting the tissue image data into at least one patch tissue image such that the coordinates of the central spot among the multiple spots become central coordinates of the rectangular box.

16 . A non-transitory computer-readable recording medium in which a computer program for executing the method for predicting cell type enrichment from tissue images according to claim 15 is recorded.

17 . A non-transitory computer-readable recording medium in which a computer program for executing the method for predicting cell type enrichment from tissue images according to claim 8 is recorded.

18 . A non-transitory computer-readable recording medium in which a computer program for executing the method for predicting cell type enrichment from tissue images according to claim 7 is recorded.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Oct 16, 2023
From: CHOI, HONGYOON; KIM, YOUNG TAE; NA, KWON JOONG
To: PORTRAI INC.
Reel/Frame 065231/0934 →
Priority Claims (2)
KR 10-2021-0049331 · Apr 15, 2021 · national
KR 10-2021-0110364 · Aug 20, 2021 · national
Continuity (1)
Related Publication 20240194292A1 · Jun 13, 2024
References Cited (26)
US 10354747B1 · DePristo · 2019 [cited by examiner]
US 10467754B1 · Ando · 2019 [cited by examiner]
US 20050165290A1 · Kotsianti et al. · 2005 [cited by applicant]
US 20180353072A1 · Gelbman · 2018 [cited by examiner]
US 20220367053A1 · Mahmood · 2022 [cited by examiner]
US 20240194292A1 · Choi · 2024 [cited by examiner]
CN 112185465A · 2021 [cited by applicant]
JP 2021032705A · 2021 [cited by applicant]
KR 1020200058662A · 2020 [cited by applicant]
KR 102108050B1 · 2020 [cited by applicant]
WO 2021067514A1 · 2021 [cited by applicant]
Alona Levy-Jurgenson et al., “Spatial transcriptomics inferred from pathology whole-slide images links tumor heterogeneity to survival in breast and lung cancer,” Scientific Reports, 2020. [cited by applicant]
The Australian Examination Report for Australian Patent Application No. 2022257481, dated Oct. 17, 2024. [cited by applicant]
Sungwoo Bae, et al., “Discovery of molecular features underlying morphological landscape by integrating spatial transcriptomic data with deep features of tissue image,” bioRxiv, 2020, 150698. [cited by applicant]
Tommaso Biancalani et al., “Deep learning and alignment of spatially-resolved whole transcriptomes of single cells in the mouse brain with Tangram,” bioRxiv, 2020, 272831. [cited by applicant]
Bryan He et al., “Integrating spatial gene expression and breast tumour morphology via deep learning,” Nature biomedical engineering, 2020, pp. 827-834, 4(8). [cited by applicant]
Floyd Maseda et al., “DEEPsc: a deep learning-based map connecting single-cell transcriptomics and spatial imaging data,” Frontiers in Genetics, 2021, vol. 12, 636743. [cited by applicant]
Xiao Tan, et al., “SpaCell: integrating tissue morphology and spatial gene expression to predict disease cells,” Bioinformatics, 2019, pp. 2293-2294, 36(7). [cited by applicant]
Ruben Dries et al., “Giotto, a toolbox for integrative analysis and visualization of spatial expression data,” Genome Biology, 2021, 22:78. [cited by applicant]
The Extended European Search Report for European Patent Application No. 22788243.8, dated Feb. 12, 2025. [cited by applicant]
Benoît Schmauch et al., “A deep learning model to predict RNA-Seq expression of tumours from whole slide images,” Nature Communications, Aug. 3, 2020, vol. 11. [cited by applicant]
Qianqian Song et al., “DSTG: deconvoluting spatial transcriptomics data through graph-based artificial intelligence”, Briefings in Bioinformatics, Jan. 22, 2021, vol. 22, Issue. 5, pp. 1-13. [cited by applicant]
Niyaz Yoosuf et al., “Identification and transfer of spatial transcriptomics signatures for cancer diagnosis”, Breast Cancer Research, 2020, vol. 22:6, pp. 1-10. [cited by applicant]
Sungwoo Bae et al., “CellDART: Cell type inference by domain adaptation of single-cell and spatial transcriptomic data”, bioRxiv, Apr. 27, 2021. [cited by applicant]
Eduard Chelebian et al., “Morphological Features Extracted by AI Associated with Spatial Transcriptomics in Prostate Cancer”, Cancers, Sep. 28, 2021, vol. 13:4837, pp. 1-14. [cited by applicant]
Sungwoo Bae et al., Discovery of molecular features underlying the morphological landscape by integrating spatial transcriptomic data with deep features of tissue images, Nucleic Acids Research, Feb. 22, 2021, vol. 49, … [cited by applicant]