Back

Signal Strength Aware Latent Spaces Reveal Molecularly Distinct Substructures within Human Kidney Tissue

Delacour, P.-L.; Migas, L.; Farrow, M.; Yang, H.; Fogo, A. B.; Spraggins, J. M.; Van de Plas, R.

2026-01-16 bioinformatics
10.64898/2026.01.15.699686 bioRxiv
Show abstract

As datasets grow increasingly high-dimensional and complex, distinguishing a condensed set of interpretable underlying factors becomes essential. In spatial omics, for example, hundreds to thousands of molecular features per observation promise unprecedented biological insight. However, without meaningful latent representations, that potential remains markedly untapped. We propose a new approach based on the beta-variational autoencoder and kernel density estimation to dissect data along independent, uncertainty-aware, and interpretable (yet non-linear) latent axes. We include a novel comparative-latent-traversal algorithm to translate latent findings back into the original measurement context. Demonstrating on imaging mass spectrometry-based molecular imaging of human kidney, the approachs disentangling properties are shown to impress a latent space structure that separates signal strength from relative signal content, offering exceptional chemical insight. Our approach uncovers unexpected subdivisions within kidney proximal tubules, confirmed to be biological, and reveals hereto-unknown lipid species differentiating them. This confirms our workflows potential as an interpretation-and-hypothesis-generating discovery tool.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

1
Bioinformatics
◐ 1204 papers in training set
Top 0.9%
22.4%
Similar papers in this journal
2
PLOS Computational Biology
● 1863 papers in training set
Top 4%
9.1%
Similar papers in this journal
3
Patterns
○ 78 papers in training set
Top 0.2%
5.6%
Similar papers in this journal
4
Nature Communications
○ 5641 papers in training set
Top 26%
5.6%
Similar papers in this journal
5
Briefings in Bioinformatics
◐ 354 papers in training set
Top 2%
4.4%
Similar papers in this journal
6
Nature Methods
○ 385 papers in training set
Top 3%
2.7%
7
Analytical Chemistry
● 218 papers in training set
Top 1%
2.7%
50% of probability mass above
8
PLOS ONE
● 5266 papers in training set
Top 40%
2.5%
9
Nature Machine Intelligence
○ 70 papers in training set
Top 1%
2.2%
10
Bioinformatics Advances
◐ 203 papers in training set
Top 3%
2.2%
11
GigaScience
◐ 212 papers in training set
Top 2%
1.8%
12
IEEE Journal of Biomedical and Health Informatics
● 37 papers in training set
Top 0.6%
1.7%
13
Genome Biology
○ 637 papers in training set
Top 6%
1.5%
14
eLife
● 5828 papers in training set
Top 54%
1.5%
15
Cell Reports Methods
○ 165 papers in training set
Top 2%
1.4%
16
Journal of Proteome Research
● 234 papers in training set
Top 1%
1.4%
17
Molecular & Cellular Proteomics
183 papers in training set
Top 1%
1.3%
18
Biological Imaging
◐ 15 papers in training set
Top 0.2%
1.2%
19
Proceedings of the National Academy of Sciences
● 2444 papers in training set
Top 33%
1.2%
20
Communications Biology
○ 993 papers in training set
Top 20%
1.2%
21
Cell Systems
○ 201 papers in training set
Top 4%
1.1%
22
NeuroImage
○ 903 papers in training set
Top 5%
1.1%
23
PNAS Nexus
◐ 159 papers in training set
Top 3%
1.0%
24
Nucleic Acids Research
◐ 1281 papers in training set
Top 12%
1.0%
25
Nature Biotechnology
○ 172 papers in training set
Top 4%
0.9%
26
Analytical and Bioanalytical Chemistry
○ 18 papers in training set
Top 0.3%
0.9%
27
Frontiers in Bioinformatics
○ 49 papers in training set
Top 1%
0.9%
28
Scientific Reports
○ 3612 papers in training set
Top 71%
0.9%
29
BMC Bioinformatics
○ 457 papers in training set
Top 5%
0.9%
30
Forensic Science International: Genetics
○ 26 papers in training set
Top 0.1%
0.9%

"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.