ISSAAC-seq is a single-cell multiomic method for simultaneous profiling of chromatin accessibility and gene expression from the same cell.
(Image created with BioRender)
The repository contains details of the library structures and the code for the data processing and analysis. There are two ways of executing the method with small differences in terms of sample preparation and data preprocessing. Go the the FACS page to check the details about the plate-based workflow. Go to the Droplet page to check the droplet-based workflow. In the droplet-based workflow, we used 10x Genomics Single Cell ATAC kit for the demonstration, but it should work on any system with a Tn5 capture sequence, such as BioRad ddSEQ and HyDrop.
The raw data (fastq files) are deposited into ArrayExpress under the accession number E-MTAB-11264.
The ISSAAC-seq experiments and other methods, including sci-CAR-seq, SNARE-seq, Paired-seq, SHARE-seq and 10x Multiome, are organised in their own subdirectories. In each subdirectory, a snakemake file ( Snakefile ) is provided to perform the data preprocessing steps. The exact programs and commands can be found there. Briefly, each Snakefile starts with downloading the raw data from the public repository, performing the required file manipulations and mapping the reads using chromap (for ATAC) or STARsolo (for RNA). Finally, it produces the cell-by-gene or cell-by-peak count matrix, together with a metrics.csv file which contains the quality information, including the numbmer of reas/UMIs and the number of detected peaks/genes for each cell.
The metrics.csv files are provided already, and the count matrices can be generated by simply run the Snakefile in each subdirctory. After that, we provide a few Jupyter Notebooks to show how exactly we perform the downstream comparisons and data analysis:
technical_qc_comparisons.ipynb shows how exactly we perform the comparisons of different methods in different biological samples.
mCortex_cell_types.ipynb shows how the cell type annotation is done.
mCortex_pseudotime.ipynb shows how the peudotime analysis is done.
For future analysis of your own ISSAAC-seq data, simply get the Snakfile from any FACS or droplet subdirectory, modify the content there, such as removeing the downsampling steps. To run the pipeline using all available cores, simply run:
snakemake --cores
- General programs
- Python3
- R v4
- Python packages
- numpy (v1.19.5)
- scipy (v1.5.3)
- pandas (v1.1.5)
- seaborn (v0.11.0.rc0)
- scanpy (v1.9.0)
- scikit-learn (v0.22.1)
- R packages
- Matrix (v1.3-3)
- Signac (v1.6.0)
- Seurat (v4.1.0)
- monocle (v2.20.0)
- GenomeInfoDb (v1.30.1)
- EnsDb.Mmusculus.v79 (v2.99.0)
- EnsDb.Hsapiens.v86 (v2.99.0)
- ggplot2 (v3.35)
- patchwork (v1.1.1)
- dplyr (v1.0.8)
- stringr (v1.4.0.9000)
- mgsub (v1.7.3)
- stringi (v1.7.6)
- Other Bioinformatics Utilities
- chromap (0.2.1-r369)
- MACS2 (v2.2.7.1)
- STARsolo (v2.7.9a)
- Cellranger-ATAC (v2.0.0)
- samtools (v1.9)
- bedtools (v2.30.0)
- seqtk (v1.3-r106)
- sratoolkit (v3.0.0)
- faSize, bedClip, calc, addCols from UCSC utilities
Xu W, Yang W, Zhang Y, Chen Y, Zhang Q, Wang X, Song K, Jin W* and Chen X* (2021) ISSAAC-seq enables sensitive and flexible multimodal profiling of chromatin accessibility and gene expression in single cells. Nat. Methods, 19, 1243–1249 (2022)
Xi Chen
[email protected]