PairPlex uses combinatorial barcoding (inspired by UDA-seq) to perform cost-effective sequencing of large numbers of natively paired antibodies by massively overloading 10x Genomics reactions.
PairPlex can be installed with pip:
pip install pairplexAlternatively, you can install from source (which may not be entirely stable, so use at your own risk):
git clone https://github.com/brineylab/pairplex
cd pairplex
pip install .Installation can be quickly confirmed by checking the version:
pairplex versionIf the current version is displayed, installation was successful.
pairplex run /path/to/sequencing_data /path/to/output_directorySequencing data can be provided as a single FASTA/Q file or a directory containing one or more FASTA/Q files (optionally gzip compressed). If a directory is provided, all files in the directory will be processed (recursively).
Providing the merge_paired_reads option will instruct PairPlex to merge paired-end sequencing reads with fastp prior to processing, so the output of Illumina's bcl2fastq or Element's bases2fastq can be used directly as input:
pairplex run --merge_paired_reads /path/to/paired_reads /path/to/output_directoryBy default, we assume Illumina-style naming conventions for paired-end read files. For paired FASTQ files produced by Element's bases2fastq, the platform option should be set to element.
pairplex run --merge_paired_reads --platform element /path/to/paired_reads /path/to/output_directoryNote
If the sequencing data is already merged (or wasn't paired-end to begin with), the platform option is not used, since the file naming conventions are only needed to match file pairs for read merging.
The complete list of CLI options can be displayed by running:
pairplex run --helpimport pairplex
pairplex.run(
sequences="/path/to/sequencing_data",
output_directory="/path/to/output_directory",
merge_paired_reads=True,
)This repo ships SimPlex (import simplex), a companion simulator for split-emulsion (overloaded 10x + well split) libraries. It generates synthetic raw reads with known ground truth and models the physical processes that cause mispairing: droplet overloading, pre-split leakage of barcoded cDNA out of cells, ambient mRNA, UMI and sequencing errors, and 5′-truncated fragments. It also runs PairPlex on the reads and scores the output: precision, cell recall, why each lost cell was lost, and precision by confidence for pooled pairing.
Use it to measure what a PairPlex change or setting does on data that resembles your libraries before changing a production run. simplex.PRINCEPS_2025 is a calibrated profile for a heavily overloaded reaction with high leakage; tune it to your own measurements. Threshold sweeps under the legacy v1 settings do not transfer to high-leakage libraries (see the SimPlex README).
For split-emulsion libraries, pairplex pair-pooled pairs heavy and light chains across all wells by barcode. It assigns each chain a home well and reports a confidence for every pair.
See simplex/README.md for what it simulates, the available knobs and their expected effects, and how to interpret the scored results.
If you are using Pairplex or the dataset generated of paired antibody sequences, please cite:
