DoubletDecon is a transcription-based doublet detection software that uses deconvolution to identify doublets using the R statistical software. We have provided a wrapper script that takes common arguments for DoubletDecon and also provide example code for you to run manually if you prefer.


This is the data that you will need to have prepare to run DoubletDecon:


  • A QC-filtered and normalized seurat object saved as an rds object ($SEURAT_RDS)

    • For example, using the Seurat Vignette

    • If you run DoubletDecon manually, you can use any data format of interest and read in with a method that works for your data.

  • Output directory ($DOUBLETDECON_OUTDIR)

Run DoubletDecon

⏱️ Expected Resource Usage

~15min using a total of 15Gb memory when using 2 thread for the full Test Dataset which contains ~20,982 droplets of 13 multiplexed donors,

You can either run DoubletDecon with the wrapper script we have provided or you can run it manually if you would prefer to alter more parameters.


Since it is hard to predict the correct rhop to use for each dataset, we typically run a range. For example: 0.6, 0.7, 0.8, 0.9, 1, and 1.1. Then we select the results that predict the number of doublets closest to the expected doublet number. You can estimate that number with our doublet calculator The rhop parameter can be set with -r or --rhop in the command below.

First, let’s assign the variables that will be used to execute each step.

Example Variable Settings

Below is an example of the variables that we can set up to be used in the command below. These are files provided as a test dataset available in the Data Preparation Documentation Please replace paths with the full path to data on your system.

singularity exec Demuxafy.sif DoubletDecon.R -o $DOUBLETDECON_OUTDIR -s $SEURAT_RDS

You can provide many other parameters as well which can be seen from running a help request:

singularity exec image DoubletDecon.R -h

usage: DoubletDecon.R [-h] -o OUT -s SEURAT_OBJECT [-g NUM_GENES] [-r RHOP]
                      [-p SPECIES] [-n NCORES] [-c REMOVECC] [-m PMF]
                      [-f HEATMAP] [-t CENTROIDS] [-d NUM_DOUBS] [-5 ONLY50]
                      [-u MIN_UNIQ]

optional arguments:
  -h, --help            show this help message and exit
  -o OUT, --out OUT     The output directory where results will be saved
  -s SEURAT_OBJECT, --seurat_object SEURAT_OBJECT
                        A QC, normalized seurat object with classifications/clusters as Idents() saved as an rds object.
  -g NUM_GENES, --num_genes NUM_GENES
                        Number of genes to use in 'Improved_Seurat_Pre_Process' function.
  -r RHOP, --rhop RHOP  rhop to use in DoubletDecon - the number of SD from the mean to identify upper limit to blacklist
  -p SPECIES, --species SPECIES
                        The species of your sample. Can be scientific species name, KEGG ID, three letter species abbreviation, or NCBI ID.
  -n NCORES, --nCores NCORES
                        The number of unique cores you would like to use to run DoubletDecon. By default, uses one less than available detected.
                        Whether to remove clusters enriched in cell cycle genes.
  -m PMF, --pmf PMF     Whether to use unique gene expression in doublet determination.
  -f HEATMAP, --heatmap HEATMAP
                        Whether to generate heatmaps.
  -t CENTROIDS, --centroids CENTROIDS
                        Whether to use centroids instead of medoids for doublet detecting.
  -d NUM_DOUBS, --num_doubs NUM_DOUBS
                        The number of doublets to simulate for each cluster pair.
  -5 ONLY50, --only50 ONLY50
                        Whether to only compute doublets as 50:50 ratio. Default is to use other ratios as well.
  -u MIN_UNIQ, --min_uniq MIN_UNIQ
                        Minimum number of unique genes to rescue a cluster identified as doublets.

DoubletDecon Results and Interpretation

After running the DoubletDecon, you will have multiple files in the $DOUBLETDECON_OUTDIR:

├── data_processed_DoubletDecon_results.txt
├── data_processed_reclust_DoubletDecon_results.txt
├── DoubletDecon_doublets_singlets.tsv
├── DoubletDecon_doublet_summary.tsv
├── DoubletDecon_results.log
├── DRS_doublet_table_DoubletDecon_results.txt
├── DRS_results_DoubletDecon_results.txt
├── Final_doublets_exp_DoubletDecon_results.txt
├── Final_doublets_groups_DoubletDecon_results.txt
├── Final_nondoublets_exp_DoubletDecon_results.txt
├── Final_nondoublets_groups_DoubletDecon_results.txt
├── groups_processed_DoubletDecon_results.txt
├── groups_processed_reclust_DoubletDecon_results.txt
├── new_PMF_results_DoubletDecon_results.txt
├── resultsreadable_synths.txt
└── Synth_doublet_info_DoubletDecon_results.txt

DoubletDecon puts most of the results in multiple separate files. However, the wrapper script and the example code has some steps to combine these results together into a single file, which will likely be the most informative output. These are the files that we think will be the most helpful for users:

  • DoubletDecon_doublet_summary.tsv

    • A summary of the number of singlets and doublets predicted by DoubletDecon.


      Droplet N





  • DoubletDecon_doublets_singlets.tsv

    • The per-barcode singlet and doublet classification from DoubletDecon.













Merging Results with Other Software Results

We have provided a script that will help merge and summarize the results from multiple softwares together. See Combine Results.


If you used the Demuxafy platform for analysis, please reference our preprint as well as DoubletDecon.