This page describes the output file structure from the cellranger multi subcommand specifically for Flex data.

Upon completion, the cellranger multi subcommand will produce an outs/ directory with the following structure:

Using the tree Linux command, the file structure looks like this:

Copy
├── config.csv
├── multi
│   ├── count
│   └── multiplexing_analysis
└── per_sample_outs
    ├── Sample1
    └── Sample2

The first section of the outputs contains the config.csv file, a duplicate of the input config CSV file. The files in the multi/ folder are generic to the entire Flex experiment, while the files in the per_sample_outs/ directory have been demultiplexed to single samples.

Within the multi/ directory, there are count and multiplexing_analysis directories. The multiplexing_analysis sub-directory is only generated in the context of multiplex Flex inputs.

Copy
└─ multi
    ├── count
    │   ├── raw_cloupe.cloupe
    │   ├── raw_feature_bc_matrix
    │   │   ├── barcodes.tsv.gz
    │   │   ├── features.tsv.gz
    │   │   └── matrix.mtx.gz
    │   ├── raw_feature_bc_matrix.h5
    │   ├── raw_molecule_info.h5
    │   ├── raw_probe_bc_matrix.h5
    │   ├── unassigned_alignments.bam
    │   └── unassigned_alignments.bam.bai
    └── multiplexing_analysis
        ├── cells_per_tag.json
        └── frp_gem_barcode_overlap.csv

The count directory contains raw files that include cell-associated and background data from all samples within an experiment:

The multiplexing_analysis directory contains:

The cells_per_tag.json file looks like this, for each Probe Barcode (e.g., BC001) the cell-associated barcodes (e.g., "AACAAGCTCCCTCAAAACTTTAGG-1", etc.) are listed below it:

Copy
{
    "BC001":[
        "AACAAGCTCCCTCAAAACTTTAGG-1",
        "AACATAGTCCCATAGCACTTTAGG-1",
        "AACCAGGTCATGGTCCACTTTAGG-1",
...
    "BC002":[
        "AAACTGTCAGGAGCAAAACGGGAA-1",
        "AAAGGGATCTAATCGTAACGGGAA-1",
        "AACCAAATCGGTCAAGAACGGGAA-1",
...
}

The frp_gem_barcode_overlap.csv file can be used to troubleshoot scenarios where, for example, two different Probe Barcodes were accidentally added to the same hybridization reaction. For multiplexed GEX + Antibody Capture experiments, overlap between incorrect BC+AB barcode pairs may result from contamination or using the same Probe Barcode for two antibody panels. An alert in the web summary will be triggered if the overlap coefficient is ≥ 60%. Please contact [email protected] if you have questions.

Copy
barcode1_id,barcode2_id,barcode1_gems,barcode2_gems,common_gems,overlap
BC009,BC010,2333,2608,81,0.034719245606515216
[ … ]
BC012,AB011,2461,4419,188,0.07639171068671272

Column descriptions:

• barcode1_id: First barcode identifier
• barcode2_id: Second barcode identifier
• barcode1_gems: Number of 10x GEM Barcodes for barcode1
• barcode2_gems: Number of 10x GEM Barcodes for barcode2
• common_gems: Number of 10x GEM Barcodes in common
• overlap: The overlap coefficient of these two barcodes (either pair of Probe Barcodes (BC), Probe Barcode (BC) and Antibody Multiplexing Barcode (AB), or AB-AB for Antibody Capture only analysis).

Overlap coefficient = $\frac{{\text{Common GEMS}}} {\text{min(Barcode 1 GEMS, Barcode 2 GEMS)}}$

The per_sample_outs directory contains sample-level files with any data associated with a valid Probe Barcode that could be assigned to a sample, including both cells and background.

Copy
├── count
│   ├── analysis
│   │   ├── clustering
│   │   ├── diffexp
│   │   ├── pca
│   │   ├── tsne
│   │   └── umap
│   ├── probe_set.csv
│   ├── sample_cloupe.cloupe
│   ├── sample_alignments.bam
│   ├── sample_alignments.bam.bai
│   ├── sample_filtered_barcodes.csv
│   ├── sample_filtered_feature_bc_matrix
│   │   ├── barcodes.tsv.gz
│   │   ├── features.tsv.gz
│   │   └── matrix.mtx.gz
│   ├── sample_filtered_feature_bc_matrix.h5
│   ├── sample_raw_feature_bc_matrix
│   │   ├── barcodes.tsv.gz
│   │   ├── features.tsv.gz
│   │   └── matrix.mtx.gz
│   ├── sample_raw_feature_bc_matrix.h5
│   ├── sample_molecule_info.h5
|   └── sample_raw_probe_bc_matrix.h5
├── metrics_summary.csv
└── web_summary.html

The per_sample_outs directory contains:

The count directory contains:

Upon completion, the cellranger multi subcommand will produce an outs/ directory with the following structure:

Copy
├── config.csv
├── multi
│   └── count
└── per_sample_outs
    └── Sample1

Within the multi/ directory, there is a count/ directory. This directory contains the same files as described above for the multiplex Flex count/ directory. There is no multiplexing_analysis/ directory for singleplex experiments.

Copy
└─ multi
   └── count
   ├── raw_cloupe.cloupe
   ├── raw_feature_bc_matrix
   │   ├── barcodes.tsv.gz
   │   ├── features.tsv.gz
   │   └── matrix.mtx.gz
   ├── raw_feature_bc_matrix.h5
   ├── raw_molecule_info.h5
   ├── raw_probe_bc_matrix.h5
   ├── unassigned_alignments.bam
  └── unassigned_alignments.bam.bai

The per_sample_outs directory contains data associated with all valid barcodes in the singleplex Flex library. This directory contains the same files as described above for the multiplex Flex per_sample_outs directory, except for the sample_raw_feature_bc_matrix/ and sample_raw_feature_bc_matrix.h5 files. The web_summary.html is described on the multi web summary page.

Copy
├── count
│   ├── analysis
│   │   ├── clustering
│   │   ├── diffexp
│   │   ├── pca
│   │   ├── tsne
│   │   └── umap
│   ├── probe_set.csv
│   ├── sample_cloupe.cloupe
│   ├── sample_alignments.bam
│   ├── sample_alignments.bam.bai
│   ├── sample_filtered_barcodes.csv
│   ├── sample_filtered_feature_bc_matrix
│   │   ├── barcodes.tsv.gz
│   │   ├── features.tsv.gz
│   │   └── matrix.mtx.gz
│   ├── sample_filtered_feature_bc_matrix.h5
│   ├── sample_molecule_info.h5
│   └── sample_raw_probe_bc_matrix.h5
├── metrics_summary.csv
└── web_summary.html

The Flex analyses with an Antibody Capture library and/or CRISPR Guide Capture library have all the same outputs described in the per_sample_outs/count directory of singleplex and multiplex GEX analyses. In addition, they contain a feature_reference.csv file, which is a duplicate of the input Feature Reference CSV file.

Copy
├── count
│   ├── analysis
│   │   ├── clustering
│   │   ├── diffexp
│   │   ├── pca
│   │   ├── tsne
│   │   └── umap
│   ├── aggregate_barcodes.csv
│   ├── feature_reference.csv
│   ├── ...

The /multi/count directory also contains a copy of the feature_reference.csv file:

Copy
└─ multi
├── count
│   ├── raw_cloupe.cloupe
│   ├── raw_feature_bc_matrix
│   │   ├── barcodes.tsv.gz
│   │   ├── features.tsv.gz
│   │   └── matrix.mtx.gz
│   ├── feature_reference.csv
│   ├── raw_feature_bc_matrix.h5
│   ├── raw_molecule_info.h5
│   ├── unassigned_alignments.bam
│   └── unassigned_alignments.bam.bai

The web_summary.html is described on the multi web summary page.