Single-cell RNA sequencing (scRNA-seq) measures gene expression at the resolution of individual cells. Massively multiplexed single-cell profiling has enabled large-scale transcriptional analyses of thousands of cells in complex tissues. In most cases, the true identity of individual cells is unknown and needs to be inferred from the transcriptomic data. Existing methods typically cluster (group) cells based on similarities of their gene expression profiles and assign the same identity to all cells within each cluster using the averaged expression levels. However, scRNA-seq experiments typically produce low-coverage sequencing data for each cell, which hinders the clustering process. We introduce scMatch, which directly annotates single cells by identifying their closest match in large reference datasets. We used this strategy to annotate various single-cell datasets and evaluated the impacts of sequencing depth, similarity metric and reference datasets. We found that scMatch can rapidly and robustly annotate single cells with comparable accuracy to another recent cell annotation tool (SingleR), but that it is quicker and can handle considerably larger reference datasets. We demonstrate how scMatch can handle large customized reference gene expression profiles that combine data from multiple sources, thus empowering researchers to identify cell populations in any complex tissue with the desired precision.
scMatch is maintained by Rui Hou [[email protected]]
git clone https://github.com/asrhou/scMatch.gitA truncated FANTOM5 reference dataset can be downloaded from https://github.com/asrhou/scMatch/tree/master/refDB/FANTOM5. The compressed files need to be decompressed before being used as the reference database. We also merged it with reference datasets from SingleR, which can be downloaded from https://figshare.com/s/efd2969ce20fae5c118f.
Carlos Biagi Jr [[email protected]] is maintaining a scMatch Docker image of scMatch in https://hub.docker.com/r/biagii/scmatch and https://github.com/cbiagii/scmatch. The Docker image has all required packages installed and contains both reference datasets mentioned above. This Docker image can greatly simplify the deployment of scMatch!
This tool provides command line utilities only for now.
This tool can annotate the given transcriptome data with sample names or ontology terms. It can analyze data from multiple platforms. The required user inputs are a raw count matrix file in CSV format, or an mtx file and two tsv files generated by CellRanger.
scMatch: Annotate the given transcriptome data using human and/or mouse expression data from given reference dataset.
scMatch.py [-h] [--refType REFTYPE] [--testType TESTTYPE] --refDS REFDS
[--dFormat DFORMAT] --testDS TESTDS
[--testMethod TESTMETHOD] [--testGenes TESTGENES]
[--keepZeros KEEPZEROS] [--coreNum CORENUM]
optional arguments:
-h, --help show this help message and exit
--refType REFTYPE human (default) | mouse | both
--testType TESTTYPE human (default) | mouse | rat | chimpanzee |zebrafish
--refDS REFDS path to the folder of reference dataset(s)
--dFormat DFORMAT 10x (default) | csv
--testDS TESTDS path to the folder of test dataset if dtype is 10x,
otherwise, the path to the file
--testMethod TESTMETHOD
s[pearman] (default) | p[earson] | both
--testGenes TESTGENES
optional, path to the csv file whose first row is the
genes used to calculate the correlation coefficient
--keepZeros KEEPZEROS
y[es] (default) | n[o]
--coreNum CORENUM number of the cores to use, default=1
The raw output, generated by scMatch is an excel file showing the correlation coefficient of each reference sample with test transcriptome.
To annotate a scRNA-seq dataset 'GSE81861_Cell_Line_COUNT.csv' from Li et al., (Nature Genetics, 2017), which were sequenced by SMARTer Ultra-Low RNA Kit and can be found at NCBI GEO:GSE81861, use the following code
python scMatch.py --refDS refDB/FANTOM5 --dFormat csv --testDS GSE81861_Cell_Line_COUNT.csvA snippet of the output for a test transcriptome looks like this:
| sample name | Spearman correlation coefficient |
|---|---|
| lung adenocarcinoma cell line:A549.CNhs11275.10499-107C4 | 0.583425488 |
| bile duct carcinoma cell line:TFK-1.CNhs11265.10496-107C1 | 0.551511709 |
| hepatoma cell line:Li-7.CNhs11271.10484-107A7 | 0.545313045 |
Note: You have to change the row names of 'GSE81861_Cell_Line_COUNT.csv' to gene symbols before you run scMatch.
toTerms.py [-h] --splF SPLF --refDS REFDS [--coreNum CORENUM]
optional arguments:
-h, --help show this help message and exit
--splF SPLF path to the original sample annotation folder which
contains original sample annotation data
--refDS REFDS path to the folder of reference dataset(s)
--coreNum CORENUM number of the cores to use, default is 1
The output of toTerms includes an excel file showing the averaged correlation coefficient of each ontology terms with test transcriptome and a csv file showing the most related ontology term for each test cell.
To transfer the annotation results of 'GSE81861_Cell_Line_COUNT.csv' to the vectors of ontology terms, use the following command
python toTerms.py --splF GSE81861_Cell_Line_COUNT/annotation_result_keep_all_genes --refDS FANTOM5A snippet of the output for a test transcriptome looks like this:
| Ont Term | Avg Score |
|---|---|
| DOID:3910 ! lung adenocarcinoma | 0.54626792 |
| DOID:4897 ! bile duct carcinoma | 0.538693164 |
| DOID:4556 ! lung large cell carcinoma | 0.532440558 |
visAnnos.py [-h] --testDS TESTDS [--dFormat DFORMAT] --annoFile
ANNOFILE [--visMethod VISMETHOD]
optional arguments:
-h, --help show this help message and exit
--testDS TESTDS path to the folder of test dataset if dtype is 10x,
otherwise, the path to the file
--dFormat DFORMAT 10x (default) | csv
--annoFile ANNOFILE path to the annotation file generated by scMatch
--visMethod VISMETHOD
p[ca] (default) | t[sne] | u[map]
If users have installed seaborn, scikit-learn, umap and plotly, then visAnnos can arrange single cells in a 2D space based on the output of PCA, tSNE, or UMAP and generated an interactive plot using plotly in an html file. The cells are colored based on the annotations in the file like: 'human_Spearman_top_ann.csv' from scMatch.py and 'human_Spearman_Avg_top_ann.csv' from toTerms.py.
To visualise the annotation results of 'GSE81861_Cell_Line_COUNT.csv' through UMAP, use the following commands
python visAnnos.py --dFormat csv --testDS GSE81861_Cell_Line_COUNT.csv --annoFile GSE81861_Cell_Line_COUNT/annotation_result_keep_all_genes/human_Spearman_top_ann.csv --visMethod uor
python visAnnos.py --dFormat csv --testDS GSE81861_Cell_Line_COUNT.csv --annoFile GSE81861_Cell_Line_COUNT/annotation_result_keep_all_genes/human_Spearman_Avg_top_ann.csv --visMethod u