Back

scJET: Full-gene Space Single-cell Expression Generation with Patch-based Transformer Modeling

Liang, Q.; Lyu, Q. R.

2026-07-09 bioinformatics
10.64898/2026.07.06.736701 bioRxiv
Show abstract

Most single-cell generative models rely on highly variable genes (HVGs) or low-dimensional latent representations, limiting their capacity to capture the complexity of full-gene features. We present scJET, a patch-based Transformer denoising framework that operates in full-gene space. scJET preserves global manifold structure, local neighborhood statistics, and gene-level expression programs. By combining scalable patch tokenization with full-gene denoising, scJET provides an efficient framework for transcriptome-wide single-cell matrix generation.

Matching journals

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

1
Nature Methods
○ 385 papers in training set
Top 0.2%
30.5%
Similar papers in this journal
2
Nature Communications
○ 5641 papers in training set
Top 14%
12.7%
Similar papers in this journal
3
Nature
○ 645 papers in training set
Top 2%
6.6%
Similar papers in this journal
4
Nature Biotechnology
○ 172 papers in training set
Top 0.6%
6.1%
Similar papers in this journal
50% of probability mass above
5
Genome Biology
○ 637 papers in training set
Top 3%
5.1%
Similar papers in this journal
6
Nature Machine Intelligence
○ 70 papers in training set
Top 0.6%
4.2%
7
Cell Systems
○ 201 papers in training set
Top 1%
4.2%
8
Nature Biomedical Engineering
○ 47 papers in training set
Top 0.3%
4.0%
9
Bioinformatics
◐ 1204 papers in training set
Top 7%
1.7%
10
Science
● 477 papers in training set
Top 5%
1.7%
11
Advanced Science
○ 286 papers in training set
Top 5%
1.7%
12
Nucleic Acids Research
◐ 1281 papers in training set
Top 9%
1.7%
13
Proceedings of the National Academy of Sciences
● 2444 papers in training set
Top 28%
1.7%
14
Nature Genetics
○ 286 papers in training set
Top 3%
1.7%
15
Cell Genomics
○ 172 papers in training set
Top 4%
1.1%
16
Nature Chemical Biology
○ 119 papers in training set
Top 2%
1.1%
17
Nature Cell Biology
○ 118 papers in training set
Top 2%
1.1%
18
Nature Computational Science
○ 55 papers in training set
Top 1%
1.0%
19
Science Advances
● 1243 papers in training set
Top 29%
1.0%
20
Briefings in Bioinformatics
◐ 354 papers in training set
Top 7%
0.9%
21
Cell Reports Methods
○ 165 papers in training set
Top 4%
0.8%
22
Genome Research
● 468 papers in training set
Top 6%
0.8%
23
Patterns
○ 78 papers in training set
Top 3%
0.6%
24
PLOS Computational Biology
● 1863 papers in training set
Top 22%
0.6%
25
Cell
○ 431 papers in training set
Top 11%
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.