Back

Tissue reassembly with generative AI

Yu, T.; Ekbote, C.; Morozov, N.; Fan, J.; Frossard, P.; d'Ascoli, S.; Brbic, M.

2025-02-17 bioinformatics
10.1101/2025.02.13.638045 bioRxiv
Show abstract

The spatial arrangement of cells is fundamental to their function, but single-cell RNA sequencing loses spatial context by dissociating cells from tissues. We present LUNA, a generative AI model that reassembles dissociated cells into tissue structures solely from gene expression by learning spatial priors from existing spatially resolved datasets. We apply and validate LUNA across multiple technologies including reconstructing the MERFISH whole mouse brain atlas with over 1.2 million cells, de novo reassembly of mouse central nervous system scRNA-seq atlas and inference of the spatial locations of nuclei lost during Slide-tags profiling. Furthermore, we show that LUNA generalizes to unseen cell types and to pathological samples in the Parkinsons disease mouse model, identifying regions of change due to pathology. We envision that AI-driven tissue reassembly can help to overcome current technological limitations and advance our understanding of tissue organization and function.

Matching journals

The top 2 journals account for 50% of the predicted probability mass.

1
Nature Methods
○ 385 papers in training set
Top 0.1%
40.4%
Similar papers in this journal
2
Nature Communications
○ 5641 papers in training set
Top 11%
15.4%
Similar papers in this journal
50% of probability mass above
3
Nature Biotechnology
○ 172 papers in training set
Top 0.1%
10.9%
Similar papers in this journal
4
Genome Biology
○ 637 papers in training set
Top 3%
4.1%
Similar papers in this journal
5
Nature
○ 645 papers in training set
Top 4%
3.3%
Similar papers in this journal
6
Science
● 477 papers in training set
Top 3%
3.3%
7
Cell
○ 431 papers in training set
Top 5%
2.2%
8
Nature Machine Intelligence
○ 70 papers in training set
Top 1%
1.8%
9
Nature Genetics
○ 286 papers in training set
Top 3%
1.8%
10
Nature Biomedical Engineering
○ 47 papers in training set
Top 0.9%
1.5%
11
Proceedings of the National Academy of Sciences
● 2444 papers in training set
Top 33%
1.2%
12
Cell Systems
○ 201 papers in training set
Top 4%
1.1%
13
Science Advances
● 1243 papers in training set
Top 27%
1.1%
14
eLife
● 5828 papers in training set
Top 63%
1.0%
15
Nature Computational Science
○ 55 papers in training set
Top 2%
0.9%
16
Nucleic Acids Research
◐ 1281 papers in training set
Top 13%
0.9%
17
Advanced Science
○ 286 papers in training set
Top 9%
0.9%
18
Cell Reports
○ 1498 papers in training set
Top 28%
0.6%
19
Bioinformatics
◐ 1204 papers in training set
Top 9%
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.