TriTan: An efficient triple non-negative matrix factorisation method for integrative analysis of single-cell multiomics datarevision.zip contains tutorial and benchmark datasets

Three multiple publicly available single-cell multi-modal (two modalities - RNA and ATAC) datasets of varying sizes to comprehensively benchmark TriTan against other methods and leverage the output from TriTan for extensive downstream analyses. 
PBMC-10K: Human peripheral blood mononuclear cells of 10X Multiome from(https://www.10xgenomics.com/resources/datasets/pbmc-from-a-healthy-donor-granulocytes-removed-through-cell-sorting-10-k-1-standard-1-0-0) .  pbmc10k.h5mu : .h5mu data with two modalities; PBMC-10K-celltype.txt: ground truth cell type labels.
Bone marrow mononuclear cells (NeurIPS 2021 ) of 10X Multiome from GSE194122. GSE194122.zip contains prepocessed .h5mu data as well as ground truth.
Mouse skin cells using SHARE-seq protocol from GSM4156597. The .zip data contains prepocessed and unprepossed scRNA-seq and scATAC-seq data.;  GSM4156597_skin_celltype.txt:ground truth cell type labels.

本研究提供了三个公开的多模态单细胞数据集（两种模态：RNA 和 ATAC），数据集规模不一，旨在全面评估 TriTan 方法与其他方法的性能，并充分利用 TriTan 的输出结果进行广泛的下游分析。PBMC-10K 数据集包含来自 10X Multiome 的人外周血单个核细胞，数据集链接为 (https://www.10xgenomics.com/resources/datasets/pbmc-from-a-healthy-donor-granulocytes-removed-through-cell-sorting-10-k-1-standard-1-0-0)。pbmc10k.h5mu 为 .h5mu 数据，包含两种模态；PBMC-10K-celltype.txt 包含细胞类型标签。来自 GSE194122 的骨髓单个核细胞（NeurIPS 2021）数据集，GSE194122.zip 包含预处理后的 .h5mu 数据以及真实标签。使用 SHARE-seq 协议从 GSM4156597 中获取的小鼠皮肤细胞数据，.zip 数据包含预处理和未预处理的全转录组测序（scRNA-seq）和全染色质可及性测序（scATAC-seq）数据；GSM4156597_skin_celltype.txt 包含细胞类型标签。