SC2Spa: a deep learning based approach to map transcriptome to spatial origins at cellular resolution

Linbu Liao; Esha Madan; António M. Palma; Hyobin Kim; Amit Kumar; Praveen Bhoopathi; Robert Winn; Jose Trevino; Paul Fisher; Cord Herbert Brakebusch; Gahyun Kim; Junil Kim; Rajan Gogna; Kyoung Jae Won

doi:10.1186/s12859-025-06173-6

SC2Spa: a deep learning based approach to map transcriptome to spatial origins at cellular resolution

Linbu Liao, Esha Madan, António M. Palma, Hyobin Kim, Amit Kumar, Praveen Bhoopathi, Robert Winn, Jose Trevino, Paul Fisher, Cord Herbert Brakebusch, Gahyun Kim, Junil Kim, Rajan Gogna^*, Kyoung Jae Won^*

^*Corresponding author for this work

Won Group

Research output: Contribution to journal › Journal article › Research › peer-review

1 Downloads (Pure)

Abstract

BACKGROUND: Understanding cellular heterogeneity within tissues hinges on knowledge of their spatial context. However, it is still challenging to accurately map cells to their spatial coordinates.

RESULTS: We present SC2Spa, a deep learning-based approach that learns intricate spatial relationships from spatial transcriptomics (ST) data. Benchmarking tests show that SC2Spa outperformed other predictors and accurately detected tissue architecture from transcriptome. SC2Spa successfully mapped single cell RNA sequencing (scRNA-seq) to Visium assay, providing an approach to enhance the resolution for low resolution ST data. Our test showed that SC2Spa performs well for various ST technologies and robust to spatial resolution. In addition, SC2Spa can suggest spatially variable genes that cannot be identified from previous approaches.

CONCLUSIONS: SC2Spa is a robust and accurate approach to provide single cells with their spatial location and identify spatially meaningful genes.

Original language	English
Article number	148
Journal	BMC Bioinformatics
Volume	26
Issue number	1
Number of pages	19
ISSN	1471-2105
DOIs	https://doi.org/10.1186/s12859-025-06173-6
Publication status	Published - 2025

Bibliographical note

Keywords

Deep Learning
Single-Cell Analysis/methods
Transcriptome
Gene Expression Profiling/methods
Sequence Analysis, RNA/methods
Animals
Humans
RNA-Seq

Access to Document

10.1186/s12859-025-06173-6Licence: CC BY

FulltextFinal published version, 2.33 MBLicence: CC BY

Cite this

@article{5fd7fc7807794a2098fe70bc8afc26ed,

title = "SC2Spa: a deep learning based approach to map transcriptome to spatial origins at cellular resolution",

abstract = "BACKGROUND: Understanding cellular heterogeneity within tissues hinges on knowledge of their spatial context. However, it is still challenging to accurately map cells to their spatial coordinates.RESULTS: We present SC2Spa, a deep learning-based approach that learns intricate spatial relationships from spatial transcriptomics (ST) data. Benchmarking tests show that SC2Spa outperformed other predictors and accurately detected tissue architecture from transcriptome. SC2Spa successfully mapped single cell RNA sequencing (scRNA-seq) to Visium assay, providing an approach to enhance the resolution for low resolution ST data. Our test showed that SC2Spa performs well for various ST technologies and robust to spatial resolution. In addition, SC2Spa can suggest spatially variable genes that cannot be identified from previous approaches.CONCLUSIONS: SC2Spa is a robust and accurate approach to provide single cells with their spatial location and identify spatially meaningful genes.",

keywords = "Deep Learning, Single-Cell Analysis/methods, Transcriptome, Gene Expression Profiling/methods, Sequence Analysis, RNA/methods, Animals, Humans, RNA-Seq",

author = "Linbu Liao and Esha Madan and Palma, {Ant{\'o}nio M.} and Hyobin Kim and Amit Kumar and Praveen Bhoopathi and Robert Winn and Jose Trevino and Paul Fisher and Brakebusch, {Cord Herbert} and Gahyun Kim and Junil Kim and Rajan Gogna and Won, {Kyoung Jae}",

note = "{\textcopyright} 2025. The Author(s).",

year = "2025",

doi = "10.1186/s12859-025-06173-6",

language = "English",

volume = "26",

journal = "BMC Bioinformatics",

issn = "1471-2105",

publisher = "BioMed Central Ltd.",

number = "1",

}

TY - JOUR

T1 - SC2Spa

T2 - a deep learning based approach to map transcriptome to spatial origins at cellular resolution

AU - Liao, Linbu

AU - Madan, Esha

AU - Palma, António M.

AU - Kim, Hyobin

AU - Kumar, Amit

AU - Bhoopathi, Praveen

AU - Winn, Robert

AU - Trevino, Jose

AU - Fisher, Paul

AU - Brakebusch, Cord Herbert

AU - Kim, Gahyun

AU - Kim, Junil

AU - Gogna, Rajan

AU - Won, Kyoung Jae

PY - 2025

Y1 - 2025

N2 - BACKGROUND: Understanding cellular heterogeneity within tissues hinges on knowledge of their spatial context. However, it is still challenging to accurately map cells to their spatial coordinates.RESULTS: We present SC2Spa, a deep learning-based approach that learns intricate spatial relationships from spatial transcriptomics (ST) data. Benchmarking tests show that SC2Spa outperformed other predictors and accurately detected tissue architecture from transcriptome. SC2Spa successfully mapped single cell RNA sequencing (scRNA-seq) to Visium assay, providing an approach to enhance the resolution for low resolution ST data. Our test showed that SC2Spa performs well for various ST technologies and robust to spatial resolution. In addition, SC2Spa can suggest spatially variable genes that cannot be identified from previous approaches.CONCLUSIONS: SC2Spa is a robust and accurate approach to provide single cells with their spatial location and identify spatially meaningful genes.

AB - BACKGROUND: Understanding cellular heterogeneity within tissues hinges on knowledge of their spatial context. However, it is still challenging to accurately map cells to their spatial coordinates.RESULTS: We present SC2Spa, a deep learning-based approach that learns intricate spatial relationships from spatial transcriptomics (ST) data. Benchmarking tests show that SC2Spa outperformed other predictors and accurately detected tissue architecture from transcriptome. SC2Spa successfully mapped single cell RNA sequencing (scRNA-seq) to Visium assay, providing an approach to enhance the resolution for low resolution ST data. Our test showed that SC2Spa performs well for various ST technologies and robust to spatial resolution. In addition, SC2Spa can suggest spatially variable genes that cannot be identified from previous approaches.CONCLUSIONS: SC2Spa is a robust and accurate approach to provide single cells with their spatial location and identify spatially meaningful genes.

KW - Deep Learning

KW - Single-Cell Analysis/methods

KW - Transcriptome

KW - Gene Expression Profiling/methods

KW - Sequence Analysis, RNA/methods

KW - Animals

KW - Humans

KW - RNA-Seq

U2 - 10.1186/s12859-025-06173-6

DO - 10.1186/s12859-025-06173-6

M3 - Journal article

C2 - 40457183

SN - 1471-2105

VL - 26

JO - BMC Bioinformatics

JF - BMC Bioinformatics

IS - 1

M1 - 148

ER -