Improving Single-Cell RNA-seq Clustering by Integrating Pathways
We designed a framework (sciPath) to study the accuracy and robustness of existing single-cell clustering by integrating pathways, including 10 state-of-art single-cell clustering methods and 4 pathway databases, pathway integration method and a complete set of evaluation indicators.
1. Datasets
Demo datasets are saved in ".//Demo_data", including scRNA-seq matrix (".//Demo_data//matrix"), pathway (".//Demo_data//pathway") and cell labels (".//Demo_data//label").
2. Package installation
The installation code of the offline packages and the online packages are saved in ".//package//package_install.R".
1. clustering_by_gene_only.R
Single cell clustering that only consider gene-level information, including (1)K-means, (2)Hierarchical, (3)Spectral, (4)DBSCAN, (5)SC3, (6)Seurat, (7)CIDR, (8)pcaReduce, (9)SOUP, (10)SNN-Cliq.
2. clustering_by_integreting_pathway.R
Single cell clustering integrating pathway-level information, including 10 clustering methods, pathway scoring method (AUCell, function "pathway_scoring") and integration method (SNF, function "integrating_pathway").
3. data_simulation_with_noise.R
The generation of three different expression profiles with noise, including "randomly set to 0" (function "simulation_0"), "gaussian noise" (function "simulation_gaussian") and "randomly amplify" (function "simulation_gaussian")
4. evaluation.R
The evaluation of accuracy and robustness of clustering methods, including ARI (function "evaluation_ARI"), NMI (function "evaluation_NMI"), MES (function "evaluation_MES") and AUC (function "evaluation_AUC").
Zhang C, Gao L, Wang B, Gao Y. Improving Single-Cell RNA-seq Clustering by Integrating Pathways. Briefings in Bioinformatics. 2021 May 3:bbab147. doi: 10.1093/bib/bbab147.