This repository includes benchmarking for clustering and community detection techniques.
- Summary of domain-agnostic clustering techniques. Method names along with their corresponding algorithmic details and computing platforms are provided.
Method Type | Algorithm | Library (function) [platform] |
---|---|---|
Partitional | k-means | stats (k-means) [R] |
Linkage | Hierarchical | cluster (agnes) [R] |
Model-based: Gaussian Mixtures | Hcmodel | mclust (hc) [R] |
Density-based | DBSCAN | dbscan (dbscan) [R] |
Subspace-based | Hddc | HDclassif (hddc) [R] |
Graph-based: Community Detection | Louvain | igraph(louvain) [R] |
Graph-based: Community Detection | Infomap | igraph(louvain) [R] |
Graph-based: Shared nearest neighbor | SNNclust | dbscan (sNNclust) [R] |
- A summary of domain-specific clustering techniques for single-cell RNA sequencing data. Method names along with their corresponding algorithmic details and computing platforms are provided.
Method | Type | Algorithm | Platform |
---|---|---|---|
Seurat | graph-based | PCA + KNN + Louivan | R |
Monocle3 | graph-based | PCA + KNN + Leiden | R |
CIDR | Linkage | PCA + hierarchical | R |
Sincell | Partition | K-mediods | R |
pcaReduce | Linkage/Partition | PCA + k-means + hierarchical | R |
SC3 | Partition | PCA + k-means | R |
RaceID3 | Partition | k-means | R |
sscClust | graph-based | Shared nearest neighbor | R |
- Summary of benchmark datasets with ground truth information used for evaluating single-cell clustering techniques (obtained from https://hemberg-lab.github.io/scRNA.seq.datasets/).
Data Set | Year | Technology | Data type | Number of Cells | Cell Populations | Features | PMID |
---|---|---|---|---|---|---|---|
Baron | 2016 | Droplet | UMI counts | 1886 | 13 | 14861 | 27667365 |
Darmanis | 2015 | SMARTer | Read counts | 466 | 9 | 21630 | 26060301 |
Deng | 2014 | Smart-Seq | Read counts | 268 | 6 | 21297 | 24408435 |
Goolam | 2016 | Smart-Seq2 | Read counts | 124 | 4 | 28147 | 27015307 |
Klein | 2015 | Droplet | UMI counts | 2717 | 4 | 24047 | 26000487 |
Li | 2017 | SMARTer | Read counts | 561 | 9 | 43055 | 28319088 |
Romanov | 2016 | Fluidigm C1 | Read counts | 2881 | 7 | 21143 | 27991900 |
Zeisel | 2015 | C1 UMI | UMI counts | 3005 | 9 | 19972 | 25700174 |