MiniAtlas
收藏魔搭社区2025-10-09 更新2025-04-26 收录
下载链接:
https://modelscope.cn/datasets/UCSC-VLAA/MiniAtlas
下载链接
链接失效反馈官方服务:
资源简介:
### Introduction
We propose the **MiniAtlas** dataset, containing more than 100,000 scATAC-seq with paired scRNA-seq as training data, across 19 tissues and 56 cell types, facilitating the training of foundation models. This dataset can be used to training single-cell multiomics fundation model.

### Subsets
This dataset is divided into four subsets to accommodate different research needs and access limitations:
1. `full_atlas_atac.h5ad` and `full_atlas_rna.h5ad` (~120k samples): full data of MiniAtlas, containing all tissues and cell types.
2. Evaluation set for different tissues: containing three tissues (Kidney, PBMC, BMMC), can be used to cell-type annotation or RNA-prediction fine-tuning and evaluation.
### Citation
If you find MiniAtlas useful for your research and applications, please cite using this BibTeX:
```
@article {Wu2025.02.05.636688,
author = {Wu, Juncheng and Wan, Changxin and Ji, Zhicheng and Zhou, Yuyin and Hou, Wenpin},
title = {EpiFoundation: A Foundation Model for Single-Cell ATAC-seq via Peak-to-Gene Alignment},
elocation-id = {2025.02.05.636688},
year = {2025},
doi = {10.1101/2025.02.05.636688},
URL = {https://www.biorxiv.org/content/early/2025/02/08/2025.02.05.636688},
eprint = {https://www.biorxiv.org/content/early/2025/02/08/2025.02.05.636688.full.pdf},
journal = {bioRxiv}
}
```
### 简介
我们提出**MiniAtlas**数据集,该数据集包含超过10万个带有配对单细胞RNA测序(scRNA-seq)数据的单细胞转座酶可及性测序(scATAC-seq)样本,所有样本均作为训练数据,涵盖19种组织与56种细胞类型,可为基础模型的训练提供支持,助力单细胞多组学基础模型的研发。

### 数据集子集
本数据集划分为四个子集,以适配不同的研究需求与使用权限限制:
1. `full_atlas_atac.h5ad` 与 `full_atlas_rna.h5ad`(约12万个样本):MiniAtlas的完整数据集,涵盖全部组织与细胞类型。
2. 多组织评测集:包含肾脏(Kidney)、外周血单核细胞(PBMC,Peripheral Blood Mononuclear Cell)、骨髓单核细胞(BMMC,Bone Marrow Mononuclear Cell)三种组织,可用于细胞类型注释或RNA预测的微调与评测。
### 引用说明
若您的研究与应用中用到了MiniAtlas数据集,请采用以下BibTeX格式进行引用:
@article {Wu2025.02.05.636688,
author = {Wu, Juncheng and Wan, Changxin and Ji, Zhicheng and Zhou, Yuyin and Hou, Wenpin},
title = {EpiFoundation: A Foundation Model for Single-Cell ATAC-seq via Peak-to-Gene Alignment},
elocation-id = {2025.02.05.636688},
year = {2025},
doi = {10.1101/2025.02.05.636688},
URL = {https://www.biorxiv.org/content/early/2025/02/08/2025.02.05.636688},
eprint = {https://www.biorxiv.org/content/early/2025/02/08/2025.02.05.636688.full.pdf},
journal = {bioRxiv}
}
提供机构:
maas
创建时间:
2025-04-21



