Back

IST-editing: Infinite spatial transcriptomic editing in a generated gigapixel mouse pup

Wu, J.; Berg, I.; Koelzer, V. H.

2023-12-23 bioinformatics
10.1101/2023.12.23.573175 bioRxiv
Show abstract

Advanced spatial transcriptomics (ST) techniques provide comprehensive insights into complex living systems across multiple scales, while simultaneously posing challenges in bioimage analysis. The spatial co-profiling of biological tissues by gigapixel whole slide images (WSI) and gene expression arrays motivates the development of innovative and efficient algorithmic approaches. Using Generative Adversarial Nets (GAN), we introduce Infinite Spatial Transcriptomic editing (IST-editing) and establish gene expression-guided editing in a generated gigapixel mouse pup. Trained with patch-wise high-plex gene expression (input) and matched image data (output), IST-editing enables the synthesis of arbitrarily large bioimages at inference, e.g., with a 106496x53248 resolution. After feeding edited gene expressions to the trained network, we model cell-, tissue- and animal-level morphological transitions in the generated mouse pup. Lastly, we discuss and evaluate editing effects on interpretable morphological features. The generated WSIs of the mouse pup and code are publicly released and accessible via https://github.com/CTPLab/IST-editing.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

1
Nature Methods
○ 385 papers in training set
Top 0.4%
18.6%
Similar papers in this journal
2
Nature Communications
○ 5641 papers in training set
Top 8%
18.6%
Similar papers in this journal
3
Nature Machine Intelligence
○ 70 papers in training set
Top 0.2%
7.9%
Similar papers in this journal
4
Bioinformatics
◐ 1204 papers in training set
Top 4%
5.5%
Similar papers in this journal
50% of probability mass above
5
Advanced Science
○ 286 papers in training set
Top 2%
4.1%
Similar papers in this journal
6
Patterns
○ 78 papers in training set
Top 0.5%
3.4%
7
Nature Biotechnology
○ 172 papers in training set
Top 1%
3.3%
8
Nature Biomedical Engineering
○ 47 papers in training set
Top 0.4%
3.1%
9
Scientific Reports
○ 3612 papers in training set
Top 36%
2.8%
10
Cell Reports Methods
○ 165 papers in training set
Top 0.8%
2.7%
11
Advanced Intelligent Systems
○ 11 papers in training set
Top 0.1%
2.7%
12
iScience
○ 1154 papers in training set
Top 10%
2.5%
13
Communications Biology
○ 993 papers in training set
Top 9%
2.4%
14
Briefings in Bioinformatics
◐ 354 papers in training set
Top 4%
2.1%
15
PLOS Computational Biology
● 1863 papers in training set
Top 14%
1.7%
16
eLife
● 5828 papers in training set
Top 56%
1.3%
17
Proceedings of the National Academy of Sciences
● 2444 papers in training set
Top 34%
1.1%
18
Genome Biology
○ 637 papers in training set
Top 7%
1.1%
19
Cell Systems
○ 201 papers in training set
Top 3%
1.1%
20
Small Methods
○ 29 papers in training set
Top 0.5%
1.0%
21
Nucleic Acids Research
◐ 1281 papers in training set
Top 13%
0.9%
22
Science Advances
● 1243 papers in training set
Top 30%
0.8%
23
Nature Computational Science
○ 55 papers in training set
Top 2%
0.8%
24
PNAS Nexus
◐ 159 papers in training set
Top 5%
0.6%
25
Medical Image Analysis
○ 35 papers in training set
Top 0.7%
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.