AsaruSim is an automated Nextflow workflow designed for simulating 10x single-cell Nanopore reads. This workflow aims to generate a gold standard dataset for the objective assessment and optimization of single-cell long-read methods.
Full documentation is avialable here.
Before starting, ensure the following tools are installed and properly set up on your system:
- Nextflow >= v24.04.4: A workflow engine for complex data pipelines. Installation guide for Nextflow.
- Docker or Singularity: Containers for packaging necessary software, ensuring reproducibility. Docker installation guide, Singularity installation guide.
- Git: Required to clone the workflow repository. Git installation guide.
Clone the AsaruSim GitHub repository:
git clone https://github.com/alihamraoui/AsaruSim.git
cd AsaruSimTo test your installation, we provide an automated script to download reference annotations and simulate a subset of human PBMC dataset run_test.sh.
bash run_test.shCustomize runs by editing the nextflow.config file and/or specifying parameters at the command line.
Here are the primary input parameters for configuring the workflow:
| Parameter | Description | Default Value |
|---|---|---|
matrix |
Path to the count matrix csv file (required) | test_data/matrix.csv |
bc_counts |
Path to the barcode count file | test_data/test_bc.csv |
transcriptome |
Path to the reference transcriptome file (required) | test_data/transcriptome.fa |
features |
Matrix feature counts | transcript_id |
gtf |
Path to transcriptom annotation .gtf file | null |
cell_types_annotation |
Path to cell type annotation .csv file | null |
Configuration for error model:
| Parameter | Description | Default Value |
|---|---|---|
trained_model |
Badread pre-trained error/Qscore model name | nanopore2023 |
badread_identity |
Comma-separated values for Badread identity parameters | "98,2,99" |
error_model |
Custom error model file (optional) | null |
qscore_model |
Custom Q-score model file (optional) | null |
build_model |
to build your own error/Qscor model | false |
fastq_model |
reference real read (.fastq) to train error model (optional) | false |
ref_genome |
reference genome .fasta file (optional) | false |
| Parameter | Description | Default Value |
|---|---|---|
amp |
Amplification factor | 1 |
outdir |
Output directory for results | "results" |
projectName |
Name of the project | "test_project" |
Configuration for running the workflow:
| Parameter | Description | Default Value |
|---|---|---|
threads |
Number of threads to use | 4 |
container |
Docker container for the workflow | 'hamraouii/wf-SLSim' |
docker.runOptions |
Docker run options to use | '-u $(id -u):$(id -g)' |
User can choose among 4 ways to simulate template reads.
- use a real count matrix
- estimated the parameter from a real count matrix to simulate synthetic count matrix
- specified by his/her own the input parameter
- a combination of the above options
We use SPARSIM tools to simulate count matrix. for more information a bout synthetic count matrix, please read SPARSIM documentaion.
A demonstration dataset to initiate this workflow is accessible on zenodo DOI : 10.5281/zenodo.12731408. This dataset is a subsample from a Nanopore run of the 10X 5k human pbmcs.
The human GRCh38 reference transcriptome, gtf annotation and fasta referance genome can be downloaded from Ensembl.
You can use the run_test.sh script to automatically download all required datasets.
nextflow run main.nf --matrix dataset/sub_pbmc_matrice.csv \
--transcriptome dataset/Homo_sapiens.GRCh38.cdna.all.fa \
--features gene_name \
--gtf dataset/GRCh38-2020-A-genes.gtf nextflow run main.nf --matrix dataset/sub_pbmc_matrice.csv \
--transcriptome dataset/Homo_sapiens.GRCh38.cdna.all.fa \
--features gene_name \
--gtf dataset/GRCh38-2020-A-genes.gtf \
--pcr_cycles 2 \
--pcr_dup_rate 0.7 \
--pcr_error_rate 0.00003 nextflow run main.nf --matrix dataset/sub_pbmc_matrice.csv \
--transcriptome dataset/Homo_sapiens.GRCh38.cdna.all.fa \
--features gene_name \
--gtf dataset/GRCh38-2020-A-genes.gtf \
--sim_celltypes true \
--cell_types_annotation dataset/sub_pbmc_cell_type.csvnextflow run main.nf --matrix Chu_param_preset \
--transcriptome datasets/Homo_sapiens.GRCh38.cdna.all.fa \
--features gene_name \
--gtf datasets/Homo_sapiens.GRCh38.112.gtfnextflow run main.nf --matrix dataset/sub_pbmc_matrice.csv \
--transcriptome dataset/Homo_sapiens.GRCh38.cdna.all.fa \
--features gene_name \
--gtf dataset/GRCh38-2020-A-genes.gtf \
--build_model true \
--fastq_model dataset/sub_pbmc_reads.fq \
--ref_genome dataset/GRCh38-2020-A-genome.fa nextflow run main.nf --matrix dataset/sub_pbmc_matrice.csv \
--transcriptome dataset/Homo_sapiens.GRCh38.cdna.all.fa \
--features gene_name \
--gtf dataset/GRCh38-2020-A-genes.gtf \
--sim_celltypes true \
--cell_types_annotation dataset/sub_pbmc_cell_type.csv \
--build_model true \
--fastq_model dataset/sub_pbmc_reads.fq \
--ref_genome dataset/GRCh38-2020-A-genome.fa \
--pcr_cycles 2 \
--pcr_dup_rate 0.7 \
--pcr_error_rate 0.00003After execution, results will be available in the specified --outdir. This includes simulated Nanopore reads .fastq, along with log files and QC report.
To clean up temporary files generated by Nextflow:
nextflow clean -f
- We would like to express our gratitude to Youyupei for the development of SLSim, which has been helpful to the
AsaruSimworkflow. - Additionally, our thanks go to the teams behind Badread and SPARSim, whose tools are integral to the
AsaruSimworkflow.
For support, please open an issue in the repository's "Issues" section. Contributions via Pull Requests are welcome. Follow the contribution guidelines specified in CONTRIBUTING.md.
AsaruSim is distributed under a specific license. Check the LICENSE file in the GitHub repository for details.
If you use AsaruSim in your research, please cite this manuscript:
Ali Hamraoui, Laurent Jourdren and Morgane Thomas-Chollier. AsaruSim: a single-cell and spatial RNA-Seq Nanopore long-reads simulation workflow. bioRxiv 2024.09.20.613625; doi: https://doi.org/10.1101/2024.09.20.613625