Souporcell

Souporcell is a genotype-free demultiplexing software that does not require you to have SNP genotypes the donors in your multiplexed capture. However, it can natively integrate SNP genotypes into the demultiplexing if you have them available for all the donors in your pool. If you don’t have the reference SNP genotypes for all the donors in your multiplexed pool, we have provided some scripts that will help identify clusters from given donors after running Souporcell without the SNP genotypes. Depending on your downstream analyses, if you have reference SNP genotypes for donors in your pool, you could also use Demuxlet, or Vireo.

One advantage that we have found immensely helpful about Souporcell is that it provides an ambient RNA estimate for the pool. This can be helpful to identify samples that may have high ambient RNA estimates early in the analysis pipeline so that it can be accounted for throughout downstream analyses.

Data

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

Required

• Common SNP genotypes vcf ($VCF)

  • While not exactly required, using common SNP genotype locations enhances accuracy

    • If you have reference SNP genotypes for individuals in your pool, you can use those

    • If you do not have reference SNP genotypes, they can be from any large population resource (i.e. 1000 Genomes or HRC)

  • Filter for common SNPs (> 5% minor allele frequency) and SNPs overlapping genes

• Barcode file ($BARCODES)

• Number of samples in pool ($N)

• Bam file ($BAM)

  • Aligned single cell reads

• Reference fasta ($FASTA)

  • that your reads were aligned to (or at least the same genome)

• Output directory ($SOUPORCELL_OUTDIR)

Run Souporcell

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.

VCF=/path/to/TestData4PipelineFull/test_dataset.vcf
BARCODES=/path/to/TestData4PipelineFull/test_dataset/outs/filtered_gene_bc_matrices/Homo_sapiens_GRCh38p10/barcodes.tsv
BAM=/path/to/test_dataset/possorted_genome_bam.bam
SOUPORCELL_OUTDIR=/path/to/output/souporcell
N=13

Run Souporcell Pipeline

⏱️ Expected Resource Usage

~37h using a total of 45Gb memory when using 5 threads for the full Test Dataset which contains ~20,982 droplets of 13 multiplexed donors,

You can run Souporcell with or without reference SNP genotypes - follow the instructions for each bellow:

Without Reference SNP GenotypesWith Reference SNP Genotypes

If you don’t have reference SNP genotypes for all of your donors, you can run souporcell with the following command, providing an appropriate thread number ($THREADS) for your system . Don’t worry if you only have reference SNP genotypes for a subset of your donors, we have a script that will correlate the cluster and reference SNP genotypes.

singularity exec Demuxafy.sif souporcell_pipeline.py -i $BAM -b $BARCODES -f $FASTA -t $THREADS -o $SOUPORCELL_OUTDIR -k $N --common_variants $VCF

HELP! It says my file/directory doesn’t exist!

If you receive an error indicating that a file or directory doesn’t exist but you are sure that it does, this is likely an issue arising from Singularity. This is easy to fix. The issue and solution are explained in detail in the Notes About Singularity Images

If Souporcell is successful, you will have these files in your $SOUPORCELL_OUTDIR:

/path/to/output/souporcell
├── alt.mtx
├── ambient_rna.txt
├── cluster_genotypes.vcf
├── clustering.done
├── clusters.err
├── clusters_tmp.tsv
├── clusters.tsv
├── common_variants_covered_tmp.vcf
├── common_variants_covered.vcf
├── consensus.done
├── depth_merged.bed
├── doublets.err
├── fastqs.done
├── minimap.err
├── ref.mtx
├── remapping.done
├── retag.err
├── retagging.done
├── souporcell_minimap_tagged_sorted.bam
├── souporcell_minimap_tagged_sorted.bam.bai
├── troublet.done
├── variants.done
└── vartrix.done

Additional details about outputs are available below in the Souporcell Results and Interpretation.

Souporcell Summary

We have provided a script that will provide a summary of the number of droplets classified as doublets, ambiguous and assigned to each cluster by Souporcell. You can run this to get a fast and easy summary of your results by providing the souporcell result file:

singularity exec Demuxafy.sif bash souporcell_summary.sh $SOUPORCELL_OUTDIR/clusters.tsv

which should print:

Classification	Assignment N
0	1441
1	980
10	1285
11	1107
12	1315
13	1529
2	1629
3	1473
4	1381
5	1360
6	1157
7	892
8	1111
9	1565
doublet	2757

or you can write the results to file:

singularity exec Demuxafy.sif bash souporcell_summary.sh $SOUPORCELL_OUTDIR/clusters.tsv > $SOUPORCELL_OUTDIR/souporcell_summary.tsv

Note

To check if these numbers are consistent with the expected doublet rate in your dataset, you can use our Doublet Estimation Calculator.

If the souporcell summary is successful, you will have this new file in your $SOUPORCELL_OUTDIR:

/path/to/output/souporcell
├── alt.mtx
├── ambient_rna.txt
├── cluster_genotypes.vcf
├── clustering.done
├── clusters.err
├── clusters_tmp.tsv
├── clusters.tsv
├── common_variants_covered_tmp.vcf
├── common_variants_covered.vcf
├── consensus.done
├── depth_merged.bed
├── doublets.err
├── fastqs.done
├── minimap.err
├── ref.mtx
├── remapping.done
├── retag.err
├── retagging.done
├── souporcell_minimap_tagged_sorted.bam
├── souporcell_summary.tsv
├── troublet.done
├── variants.done
└── vartrix.done

Additional details about outputs are available below in the Souporcell Results and Interpretation.

Correlating Cluster to Donor Reference SNP Genotypes (optional)

If you have reference SNP genotypes for some or all of the donors in your pool, you can identify which cluster is best correlated with each donor in your reference SNP genotypes. We have provided a script that will do this and provide a heatmap correlation figure and the predicted individual that should be assigned for each cluster. You can either run it with the script by providing the reference SNP genotypes ($VCF), the cluster SNP genotypes ($SOUPORCELL_OUTDIR/cluster_genotypes.vcf) and the output directory ($SOUPORCELL_OUTDIR) You can run this script with:

Note

In order to do this, your $VCF must be reference SNP genotypes for the individuals in the pool and cannot be a general vcf with common SNP genotype locations from 1000 Genomes or HRC.

With ScriptRun in R

singularity exec Demuxafy.sif Assign_Indiv_by_Geno.R -r $VCF -c $SOUPORCELL_OUTDIR/cluster_genotypes.vcf -o $SOUPORCELL_OUTDIR

To see the parameter help menu, type:

singularity exec Demuxafy.sif Assign_Indiv_by_Geno.R -h

Which will print:

usage: Assign_Indiv_by_Geno.R [-h] -r REFERENCE_VCF -c CLUSTER_VCF -o OUTDIR

optional arguments:
-h, --help            show this help message and exit
-r REFERENCE_VCF, --reference_vcf REFERENCE_VCF
                                                The output directory where results will be saved
-c CLUSTER_VCF, --cluster_vcf CLUSTER_VCF
                                                A QC, normalized seurat object with
                                                classifications/clusters as Idents().
-o OUTDIR, --outdir OUTDIR
                                                Number of genes to use in
                                                'Improved_Seurat_Pre_Process' function.

After correlating the cluster and reference donor SNP genotypes, you should have the new results in your directory:

If the souporcell summary is successful, you will have this new file in your $SOUPORCELL_OUTDIR:

/path/to/output/souporcell
├── alt.mtx
├── ambient_rna.txt
├── cluster_genotypes.vcf
├── clustering.done
├── clusters.err
├── clusters_tmp.tsv
├── clusters.tsv
├── common_variants_covered_tmp.vcf
├── common_variants_covered.vcf
├── consensus.done
├── depth_merged.bed
├── doublets.err
├── fastqs.done
├── Genotype_ID_key.txt
├── Individual_genotypes_subset.vcf.gz
├── minimap.err
├── ref.mtx
├── ref_clust_pearson_correlation.png
├── ref_clust_pearson_correlations.tsv
├── remapping.done
├── retag.err
├── retagging.done
├── souporcell_minimap_tagged_sorted.bam
├── souporcell_summary.tsv
├── troublet.done
├── variants.done
└── vartrix.done

Additional details about outputs are available below in the Souporcell Results and Interpretation.

Souporcell Results and Interpretation

After running the Souporcell steps and summarizing the results, you will have a number of files from some of the intermediary steps. These are the files that most users will find the most informative:

• To check if these numbers are consistent with the expected doublet rate in your dataset, you can use our Expected Doublet Estimation Calculator.

• clusters.tsv

  • The Souporcell droplet classifications with the log probabilities of each donor and doublet vs singlet.

    barcode	status	assignment	log_prob_singleton	log_prob_doublet	cluster0	cluster1	cluster2	cluster3	cluster4	cluster5	cluster6	cluster7	cluster8	cluster9	cluster10	cluster11	cluster12	cluster13
    AAACCTGAGATAGCAT-1	singlet	6	-47.4906809612613	-67.16353115825044	-189.38489711217204	-167.22863078578243	-175.6243866125455	-195.88836978493757	-147.1278571646738	-162.71464140958287	-47.4906809612613	-147.57558470556503	-142.24543450475267	-137.94217556189426	-171.6924681433834	-192.9070590872178	-162.2042834302814	-141.9657291979218
    AAACCTGAGCAGCGTA-1	singlet	11	-102.80051804401324	-158.38006105671326	-357.5113573904763	-403.04676141772245	-465.3312627534814	-368.72445203224066	-362.5022337777086	-377.5322002577741	-400.12257643517944	-436.7935123280712	-364.36305907429954	-434.8878131790703	-393.42953156344277	-102.80051804401324	-369.5775718688619	-403.83637627549155
    AAACCTGAGCGATGAC-1	singlet	5	-39.97694257579923	-53.76617956926222	-135.58935896223636	-129.29863536547518	-122.20920829636167	-99.54420652897485	-139.8403265674046	-39.97694257579923	-136.5313839118704	-139.57805752070823	-113.63185227373309	-117.89083888468238	-126.95555633151154	-167.2476854256994	-127.05455963457722	-123.63808626520557
    AAACCTGAGCGTAGTG-1	singlet	3	-66.73447359908208	-79.59130566934348	-146.47954690347862	-197.54291944344263	-211.47148694945332	-66.73447359908208	-163.94180016636983	-173.4754549428176	-183.73592914945144	-163.7126225130574	-172.5171380662907	-231.65011940831332	-197.42816500995383	-167.68988627905136	-165.7006532267023	-174.74052654720117
    …	…		…	…	…	…	…	…	…	…	…	…	…	…	…	…	…	…

• ambient_rna.txt

  • The estimated ambient RNA percent in the pool. We typically see < 5% for scRNA-seq PBMCs and < 10% for other scRNA-seq cell types.

    ambient RNA estimated as 4.071468697320357%
    

If you ran the Assign_Indiv_by_Geno.R script, you will also have the following files:

• Genotype_ID_key.txt

  • Key of the cluster and assignments for each individual and the Pearson correlation coefficient.

    Genotype_ID	Cluster_ID	Correlation
    113_113	5	0.9365902
    349_350	3	0.9484794
    352_353	2	0.9385500
    39_39	12	0.9325007
    40_40	8	0.9252865
    41_41	6	0.9282633
    42_42	0	0.9387788
    43_43	9	0.9497327
    465_466	11	0.9234109
    596_597	10	0.9277824
    597_598	13	0.9435752
    632_633	7	0.9179054
    633_634	1	0.9222734
    660_661	4	0.9368751

• ref_clust_pearson_correlation.png

  • Figure of the Pearson correlation coefficients for each cluster-individual pair.

    

• ref_clust_pearson_correlations.tsv

  • All of the Pearson correlation coefficients between the clusters and the individuals

    Cluster	113_113	349_350	352_353	39_39	40_40	…
    0	0.4578087241392215	0.4589573335017816	0.46351292453350446	0.48926720614880104	0.4841871441103791	…
    1	0.45706434043842825	0.48280445273461425	0.4702618797322548	0.4678187806965093	0.4801164797099736	…
    2	0.4760176832308062	0.45281488606508186	0.9385500036660724	0.47703829279476667	0.47639771569917855	…
    3	0.4771709808299328	0.9484794352067363	0.4598361363766827	0.4698832593827229	0.4822779587579728	…
    4	0.4851872933346752	0.48480637867431775	0.4908275654324142	0.48900594491809124	0.4647100675599844	…
    …	…	…	…	…	…	…

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.

Citation

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