Back

STdGCN: accurate cell-type deconvolution using graph convolutional networks in spatial transcriptomic data

Li, Y.; Luo, Y.

2023-03-12 bioinformatics
10.1101/2023.03.10.532112 bioRxiv
Show abstract

Spatially resolved transcriptomics performs high-throughput measurement of transcriptomes while preserving spatial information about the cellular organizations. However, many spatially resolved transcriptomic technologies can only distinguish spots consisting of a mixture of cells instead of working at single-cell resolution. Here, we present STdGCN, a graph neural network model designed for cell type deconvolution of spatial transcriptomic (ST) data that can leverage abundant single-cell RNA sequencing (scRNA-seq) data as reference. STdGCN is the first model incorporating the expression profiles from single cell data as well as the spatial localization information from the ST data for cell type deconvolution. Extensive benchmarking experiments on multiple ST datasets showed that STdGCN outperformed 14 published state-of-the-art models. Applied to a human breast cancer Visium dataset, STdGCN discerned spatial distributions between stroma, lymphocytes and cancer cells for tumor microenvironment dissection. In a human heart ST dataset, STdGCN detected the changes of potential endothelial-cardiomyocyte communications during tissue development.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

1
Nature Communications
○ 5641 papers in training set
Top 13%
14.7%
Similar papers in this journal
2
Briefings in Bioinformatics
◐ 354 papers in training set
Top 0.5%
10.7%
Similar papers in this journal
3
Advanced Science
○ 286 papers in training set
Top 0.4%
9.5%
Similar papers in this journal
4
Genome Biology
○ 637 papers in training set
Top 2%
7.7%
Similar papers in this journal
5
Genomics, Proteomics & Bioinformatics
188 papers in training set
Top 0.4%
5.4%
6
Nature Machine Intelligence
○ 70 papers in training set
Top 0.6%
4.7%
50% of probability mass above
7
Bioinformatics
◐ 1204 papers in training set
Top 4%
4.7%
8
PLOS Computational Biology
● 1863 papers in training set
Top 8%
4.7%
9
iScience
○ 1154 papers in training set
Top 5%
3.9%
10
Science China Life Sciences
○ 29 papers in training set
Top 0.1%
3.4%
11
Communications Biology
○ 993 papers in training set
Top 9%
2.4%
12
Cell Systems
○ 201 papers in training set
Top 2%
2.1%
13
Patterns
○ 78 papers in training set
Top 1%
2.1%
14
Small Methods
○ 29 papers in training set
Top 0.2%
1.9%
15
Cell Reports Methods
○ 165 papers in training set
Top 2%
1.5%
16
Journal of Genetics and Genomics
○ 38 papers in training set
Top 0.4%
1.3%
17
Scientific Reports
○ 3612 papers in training set
Top 65%
1.1%
18
eLife
● 5828 papers in training set
Top 60%
1.1%
19
Genome Research
● 468 papers in training set
Top 5%
1.0%
20
Science Bulletin
○ 21 papers in training set
Top 0.4%
0.8%
21
Nucleic Acids Research
◐ 1281 papers in training set
Top 14%
0.8%
22
npj Systems Biology and Applications
○ 125 papers in training set
Top 2%
0.8%
23
Cell Genomics
○ 172 papers in training set
Top 5%
0.8%
24
Nature Computational Science
○ 55 papers in training set
Top 2%
0.6%
25
Genome Medicine
○ 183 papers in training set
Top 6%
0.6%
26
Science Advances
● 1243 papers in training set
Top 34%
0.6%

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.