Many experiments generate data for multiple samples. Depending on the experimental design, these could be replicates from the same set of cells, cells from different tissues or time points from the same individual, or cells from different individuals. Samples could be processed through different Gel Bead-in Emulsion (GEM) wells or multiplexed within the same GEM well on Chromium instruments. The cellranger aggr pipeline can be used to aggregate samples from these scenarios into a single feature-barcode matrix.

The first step is to run cellranger count or cellranger multi on each individual GEM well prepared using the 10x Genomics Chromium Controller or Chromium X Series.

For example, suppose you ran three count pipelines as follows:

Copy
cd /opt/runs
cellranger count --id=LV123 ...
... wait for pipeline to finish ...
cellranger count --id=LB456 ...
... wait for pipeline to finish ...
cellranger count --id=LP789 ...
... wait for pipeline to finish ...

These three runs can be aggregated with cellranger aggr to get a single feature-barcode matrix and secondary analysis outputs. In order to do so, next create an aggregation CSV file.

Create a CSV file with a header line containing the following columns:

• sample_id: Unique identifier for this input GEM well. This will be used for labeling purposes only; it doesn't need to match any previous ID assigned to the GEM well.

• molecule_h5: Path to the molecule_info.h5 or sample_molecule_info.h5 file produced by cellranger count or multi. For example, if you processed your GEM well by calling cellranger count --id=ID in some directory /DIR, this path would be /DIR/ID/outs/molecule_info.h5.

Either make the CSV file in a text editor, or create it in Excel and save as a CSV file. Continuing with the example from the previous section, the CSV would look like this:

Copy

sample_id,molecule_h5

LV123,/opt/runs/LV123/outs/molecule_info.h5

LB456,/opt/runs/LB456/outs/molecule_info.h5

LP789,/opt/runs/LP789/outs/molecule_info.h5

In addition to the CSV columns expected by cellranger aggr, you may optionally supply columns containing library metadata (e.g., lab or sample origin). These custom library annotations do not affect the analysis pipeline but can be visualized downstream in the Loupe Browser (see Categories section below). Note that unlike other CSV inputs to Cell Ranger, these custom columns may contain characters outside the ASCII range (e.g., non-Latin characters).

For large studies involving samples run on multiple GEM wells, run cellranger count on FASTQ data from each of the GEM wells individually, and then pool the results using cellranger aggr.

The cellranger aggr command takes an aggregation CSV file specifying the list of cellranger count output files (specifically the molecule_info.h5 from each run), and produces a single feature-barcode matrix containing all the data.

When combining multiple GEM wells, the barcode sequences for each channel are distinguished by a GEM well suffix appended to the barcode sequence (see GEM wells).

By default, reads from each GEM well are subsampled such that all GEM wells have the same effective sequencing depth, measured in terms of reads that are confidently mapped to the transcriptome or assigned to the feature IDs per cell. However, it is possible to change the depth normalization mode (see Depth Normalization).

If Feature Barcode analysis is included, the input Feature Reference CSV file should be the same for each GEM well.

Starting from Cell Ranger v7.0, CRISPR Guide Capture libraries from multiple GEM wells can be aggregated with cellranger aggr. There are no changes to aggr inputs – the presence of CRISPR Guide Capture library information in the molecule_info.h5 input files enables CRISPR aggregation. Normalization is enabled by default for both Gene Expression and CRISPR Guide Capture libraries; changes to normalization parameters affect both libraries. Note that protospacer calling is performed again on the combined data included in the cellranger aggr run. CRISPR aggregation generates the crispr_analysis/ folder in the outs/ directory. The structure of the crispr_analysis folder is similar to the CRISPR outputs from cellranger count.

The cellranger aggr command takes an aggregation CSV file specifying the list of cellranger multi per sample sample_molecule_info.h5 files, and performs aggregation on any combination of Gene Expression and Feature Barcode (Antibody or CRISPR Guide Capture, Cell Multiplexing) data that are present in the individual sample outputs.

Consider two per sample datasets containing data from one 3' Cell Multiplexing experiment:

Copy
cd /opt/runs
cellranger multi --id=Run1 ...
... wait for pipeline to finish ...

To aggregate the datasets, you need to create a CSV containing the following columns:

Apart from the change in the path to the per sample sample_molecule_info.h5 file, the documentation on additional columns for creating categories, depth normalization, and batch correction are the same.

Either make the CSV file in a text editor, or create it in Excel and save as a CSV file. The CSV should look like this:

Copy
sample_id,molecule_h5
Sample1,/opt/runs/Run1/outs/per_sample_outs/Sample1/count/sample_molecule_info.h5
Sample2,/opt/runs/Run1/outs/per_sample_outs/Sample2/count/sample_molecule_info.h5

These are the most common command line arguments (run cellranger aggr --help for a full list):

After specifying these input arguments, run cellranger aggr:

Copy
cd /home/jdoe/runs
cellranger aggr --id=AGG123 \
          --csv=AGG123_libraries.csv

The pipeline will begin to run, creating a new output folder named with the specified aggregation ID (e.g. /home/jdoe/runs/AGG123). If this folder already exists, Cell Ranger will assume it is an existing pipestance and attempt to resume running it

The cellranger aggr pipeline generates output files that contain all of the data from the individual input runs, aggregated into single output files, for convenient multi-sample analysis. Refer to the Understanding Outputs section to learn about aggr output files.

When combining multiple samples into a single dataset with the cellranger aggr pipeline, you can assign categories and values to individual samples by adding columns to the cellranger aggr input spreadsheet. These category assignments propagate into Loupe Browser, where you can view them in Categories Mode to help determine genes that are differentially expressed between samples. Any columns in addition to sample_id and molecule_h5 will be converted into categories, and the cells in each sample will be assigned to one of the values in that category.

If you are aggregating libraries generated by different chemistry versions of the Single Cell Gene Expression Reagents, you might observe systematic differences in gene expression profiles between libraries. The cellranger aggr pipeline incorporates batch effect correction (algorithm details) to overcome this. To enable batch correction, include the following column in your aggregation CSV file:

• batch: Optional. Unique identifier for the batch that this GEM well belongs to. Libraries with the same batch identifier are considered to be in the same batch.

For example, if the LV123 sample in the previous example is a v2 library and the LB456 and LP789 samples are v3 libraries, set up the aggregation CSV file like this:

Copy
sample_id,molecule_h5,batch
LV123,/opt/runs/LV123/outs/molecule_info.h5,v2_lib
LB456,/opt/runs/LB456/outs/molecule_info.h5,v3_lib
LP789,/opt/runs/LP789/outs/molecule_info.h5,v3_lib

The v2_lib and v3_lib identifiers are example identifiers. Every sample from a given batch must have the same batch identifier, but otherwise the identifier text itself is arbitrary.

About chemistry batch correction

• This Chemistry Batch Correction algorithm is specifically intended to correct for systematic variability in gene expression profiles caused by different versions of the Single Cell Gene Expression chemistry. 10x Genomics has tested and verified its effectiveness primarily on aggregating Single Cell Gene Expression v2 and v3 chemistry with well-matched input material. The algorithm may be useful in other scenarios, but will require careful validation of results.

• Chemistry batch correction affects the PCA, t-SNE, and UMAP visualization and clustering results. Values in the aggregated feature-barcode matrix are not adjusted by chemistry batch correction. Differential expression analysis is still performed on the feature-barcode count matrix.

• When chemistry batch correction is enabled, FBPCA is used to perform dimensionality reduction instead of IRLBA PCA.

• The batch effect score (described in algorithm details) is recommended to compare the performance before and after batch correction. Besides the batch effect, the score also depends on the composition of the cell populations across batches.

• The minimum System Requirements of 64GB RAM will allow batch correction on datasets with a total number of 128k cells.

• Libraries with different throughput but the same chemistry can be aggregated without chemistry batch correction. For example, Single Cell 3' Gene Expression v3.1 standard kit (SK), HT, and LT can be aggregated without batch correction, as can Single Cell 5’ Gene Expression SK and HT.

Aggregating Single Cell 5' and 3' Gene Expression data

The cellranger aggr pipeline uses Chemistry Batch Correction when aggregating results from a combination of 5' and 3', or 3' v2 and 3' v3 Gene Expression data. Enabling Chemistry Batch Correction in this scenario improves the mixing of the batches in the t-SNE visualization and clustering results. We recommend using Chemistry Batch correction for these scenarios, however, residual batch effects may still be present and careful validation of the results is advised. In particular, for the V(D)J genes, the 5' Gene Expression assay will generally count the V gene segments of the immune receptor (e.g. TRBV12-1 or IGH4-2), while the 3' Gene Expression assay will count the C gene segments (e.g. TRBC or IGHA), which may pose additional analysis challenges.

The cellranger aggr pipeline supports the aggregation of Cell Multiplexing v3.1 data with Single Cell 3’ Gene Expression v3.1 data (i.e. "non-Cell Multiplexing data").

To combine Cell Multiplexing with non-Cell Multiplexing data, cellranger aggr (v6.0 and later) requires identical references (including feature/CMO reference). Therefore, non-Cell Multiplexing data must be re-run with the CMO reference file that specified Cell Multiplexing tags (CMOs) for the Cell Multiplexing data. These datasets can be run with either cellranger count (v3.0 and later) or cellranger multi (v5.0 and later). We recommend using the same version of Cell Ranger to generate inputs for cellranger aggr.

• Option 1) For cellranger count, input the CMO reference with --feature-ref and use the --no-libraries option. The cellranger count pipeline will generate a molecule_info.h5, which can be used as input to the cellranger aggr pipeline. An example of the command is below (customize the relevant file paths):

Copy

cellranger count --id=pbmc_1k_count \
         --transcriptome=/path/to/transcriptome/GRCh38-2020-A \
         --fastqs=/path/togex/fastqs/pbmc_1k_v3_fastqs/ \
         --sample=pbmc_1k_v3 \
         --feature-ref=/path/to/cmo-reference/cmo-ref.csv \
         --no-libraries

• Option 2) For cellranger multi, input the CMO feature reference in the [feature] section of the multi config CSV file. An example of the config file is below (customize the relevant file paths):

Copy

  [gene-expression]
  reference,/path/to/transcriptome/GRCh38-2020-A

  [feature]
  reference,/path/to/cmo-reference/cmo-ref.csv

  [libraries]
  fastq_id,fastqs,feature_types
  pbmc_1k_v3,/path/togex/fastqs/pbmc_1k_v3_fastqs/,Gene Expression

The cellranger multi pipeline generates a per sample sample_molecule_info.h5 file which is equivalent to the molecule_info.h5 file from a cellranger count run. These files can be used as input to the cellranger aggr pipeline. An example of the multi pipeline command is below (customize the relevant file paths):

Copy

cellranger multi --id=pbmc_1k_multi --csv=/path/to/config.csv

The Cell Multiplexing protocol has additional wash steps that may lead to depletion of more ambient mRNA from samples compared to non-Cell Multiplexing samples. This difference in mRNA may then introduce small batch effects between Cell Multiplexing and non-Cell Multiplexing datasets. However, observations from internal data show fairly small batch effects and batch effect correction is not needed in most cases. If the results for each sample reveals obvious differences between the Cell Multiplexing vs. non-Cell Multiplexing data, enabling chemistry batch correction may improve the mixing of the batches in the t-SNE visualization and clustering results.

If you are aggregating Single Cell 3' Gene Expression with Cell Multiplexing v3.1 samples with other chemistries (Single Cell 3’ v2, 5’ v2, etc.), the data can be treated similarly to Single Cell 3' v3.1 chemistry. For example, we recommend using chemistry batch correction when aggregating Single Cell 3’ v3.1 with v2 chemistry (see this article for more information about aggregating data from different chemistries).

When combining data from multiple GEM wells, the cellranger aggr pipeline automatically equalizes the average read depth per cell between groups before merging. This approach avoids artifacts that may be introduced due to differences in sequencing depth. It is possible to turn off normalization or change the way normalization is done. The none option may be appropriate if you want to maximize sensitivity and plan to handle depth normalization in a downstream step.

There are two normalization modes:

• none: Do not normalize at all.

• mapped (Default): For each library type, subsample reads from higher-depth GEM wells until they all have, on average, an equal number of reads per cell that are confidently mapped to the transcriptome (Gene Expression) or assigned to known features (Feature Barcode Technology).

Each GEM well is a physically distinct set of GEM partitions, but draws barcode sequences randomly from the pool of valid barcodes, known as the barcode whitelist. To keep the barcodes unique when aggregating multiple libraries, we append a small integer identifying the GEM well to the barcode nucleotide sequence, and use that nucleotide sequence plus ID as the unique identifier in the feature-barcode matrix. For example, AGACCATTGAGACTTA-1 and AGACCATTGAGACTTA-2 are distinct cell barcodes from different GEM wells, despite having the same barcode nucleotide sequence.

This number, which indicates which GEM well the barcode sequence came from, is called the GEM well suffix. The numbering of the GEM wells will reflect the order that the GEM wells were provided in the Aggregation CSV.