Introduction

nf-core/airrflow is a bioinformatics best-practice pipeline to analyze B-cell receptor (BCR) or T-cell receptor (TCR) repertoire sequencing data. It allows the processing of targeted bulk and single-cell adaptive immune receptor sequencing data (AIRR-seq), as well as the extraction of TCR and BCR sequences from untargeted bulk and single-cell RNA-seq data. The pipeline enables an end-to-end analysis, departing from raw reads or assembled sequences, and performs sequence assembly, V(D)J assignment, novel allele identification, genotype inference, clonal inference, repertoire analysis, lineage reconstruction and BCR/TCR sequence embedding using the Immcantation framework, as well as other immune repertoire analysis tools.

The pipeline is built using Nextflow, a workflow tool to run tasks across multiple compute infrastructures in a very portable manner. It uses Docker/Singularity containers making installation trivial and results highly reproducible. The Nextflow DSL2 implementation of this pipeline uses one container per process which makes it much easier to maintain and update software dependencies. Where possible, these processes have been submitted to and installed from nf-core/modules in order to make them available to all nf-core pipelines, and to everyone within the Nextflow community!

On release, automated continuous integration tests run the pipeline on a full-sized dataset on the AWS cloud infrastructure. This ensures that the pipeline runs on AWS, has sensible resource allocation defaults set to run on real-world datasets, and permits the persistent storage of results to benchmark between pipeline releases and other analysis sources. The results obtained from the full-sized test can be viewed on the nf-core website.

Pipeline summary

nf-core/airrflow allows the end-to-end processing of BCR and TCR bulk and single cell targeted sequencing data, as well as extracting BCR and TCR sequences from bulk and single-cell RNA-seq datasets. Several protocols are supported, please see the usage documentation for more details on the supported protocols. The pipeline has been certified as AIRR compliant by the AIRR community, which means that it is compatible with downstream analysis tools that also support this format.

1. QC and sequence assembly

• Bulk
  • Raw read quality control, adapter trimming and clipping (Fastp).
  • Filter sequences by base quality (pRESTO FilterSeq).
  • Mask amplicon primers (pRESTO MaskPrimers).
  • Pair read mates (pRESTO PairSeq).
  • For UMI-based sequencing:
    • Cluster sequences according to similarity (optional for insufficient UMI diversity) (pRESTO ClusterSets).
    • Build consensus of sequences with the same UMI barcode (pRESTO BuildConsensus).
  • Assemble R1 and R2 read mates (pRESTO AssemblePairs).
  • Remove and annotate read duplicates (pRESTO CollapseSeq).
  • Filter out sequences that do not have at least 2 duplicates (pRESTO SplitSeq).
• Single cell
  • cellranger vdj
    • Assemble contigs
    • Annotate contigs
    • Call cells
    • Generate clonotypes

1. V(D)J annotation and filtering (bulk and single-cell)

• Assign gene segments with IgBlast using a germline reference (Change-O AssignGenes).
• Annotate alignments in AIRR format (Change-O MakeDB)
• Filter by alignment quality (locus matching v_call chain, min 200 informative positions, max 10% N nucleotides)
• Filter productive sequences (Change-O ParseDB split)
• Filter junction length multiple of 3
• Annotate metadata (EnchantR)

1. QC filtering (bulk and single-cell)

• Bulk sequencing filtering:
  • Remove chimeric sequences (optional) (SHazaM, EnchantR)
  • Detect cross-contamination (optional) (EnchantR)
  • Collapse duplicates (Alakazam, EnchantR)
• Single-cell QC filtering (EnchantR)
  • Remove cells without heavy chains.
  • Remove cells with multiple heavy chains.
  • Remove sequences in different samples that share the same cell_id and nucleotide sequence, and thus are very likely contaminants.
  • Modify cell_ids to ensure they are unique in each run.

1. Novel allele and genotyping (optional, only for BCR)

• Novel allele detection (TIgGER).
• Individual genotype inference (TIgGER).
• Allele calls correction (TIgGER).

If you want to detect novel alleles and infer genotype of each individual, add the --genotyping flag in your command.

1. Clonal analysis (bulk and single-cell)

• Find threshold for clone definition (SHazaM, EnchantR).
• Create germlines and define clones (SCOPer, EnchantR)
• Repertoire analysis (EnchantR).
• Build lineage trees (Dowser, IgphyML, RAxML, EnchantR).

1. Other reporting

• Custom repertoire analysis pipeline report (Alakazam).
• Aggregate QC reports (MultiQC).

1. BCR/TCR embedding

• Translate BCR / TCR sequences with IgBLAST.
• Embed sequences with pre-trained models using AMULETY.

See the usage documentation and the parameter documentation for more details on how to use the pipeline and all the available parameters.

Usage

First, ensure that the pipeline tests run on your infrastructure:

nextflow run nf-core/airrflow -profile test,<docker/singularity/apptainer/podman/shifter/charliecloud/conda/institute> --outdir <OUTDIR>

To run nf-core/airrflow with your data, you will need to first prepare a tab-separated samplesheet with the paths to your input data and necessary metadata to run the analysis. Depending on the input data type (bulk or single-cell, raw reads or assembled reads) the input samplesheet will vary. Please follow the documentation on samplesheets for more details. An example samplesheet for running the pipeline on bulk BCR / TCR sequencing data departing from raw reads looks as follows:

Each row represents a sample with fastq files (paired-end).

A typical command to run the pipeline from bulk raw fastq files is:

nextflow run nf-core/airrflow \
-r <release> \
-profile <docker/singularity/podman/shifter/charliecloud/conda/institute> \
--mode fastq \
--input input_samplesheet.tsv \
--library_generation_method specific_pcr_umi \
--cprimers CPrimers.fasta \
--vprimers VPrimers.fasta \
--umi_length 12 \
--umi_position R1 \
--outdir ./results

For common bulk sequencing protocols we provide pre-set profiles that specify primers, UMI length, etc for common commercially available sequencing protocols. Please check the Supported protocol profiles for a full list of available profiles. An example command running the NEBNext UMI protocol profile with docker containers is:

nextflow run nf-core/airrflow \
-profile nebnext_umi,docker \
--mode fastq \
--input input_samplesheet.tsv \
--outdir results

A typical command to run the pipeline from single cell raw fastq files (10X genomics) is:

nextflow run nf-core/airrflow -r dev \
-profile <docker/singularity/podman/shifter/charliecloud/conda/institute> \
--mode fastq \
--input input_samplesheet.tsv \
--library_generation_method sc_10x_genomics \
--reference_10x reference/refdata-cellranger-vdj-GRCh38-alts-ensembl-5.0.0.tar.gz \
--outdir ./results

A typical command to run the pipeline from single-cell AIRR rearrangement tables or assembled bulk sequencing fasta data is:

nextflow run nf-core/airrflow \
-r <release> \
-profile <docker/singularity/podman/shifter/charliecloud/conda/institute> \
--input input_samp