`barbac` provides an R-native pipeline for tracking DNA barcode lineages from raw sequencing reads to time-resolved lineage trajectories. It bundles command-line pre-processing (FastQC, PEAR, minimap2, samtools) via [run_cli_pipeline()] with a native R clustering routine [super_cluster2()] and time-series visualisation [barbac_ts_area()].
Core clustering
[super_cluster2()] implements abundance-ranked greedy Levenshtein clustering with a distance-aware count-ratio merge guard, backed by a bit-parallel 64-bit edit-distance kernel written in C++. On the Johnson et al. (2023) benchmark it reaches statistical parity with Shepherd at roughly 1.7-fold the speed; on datasets with indel errors it substantially outperforms Shepherd on false-positive and wrong-sequence rates (see the package vignette).
Pipeline
The full FASTQ-to-lineage workflow is orchestrated by [run_cli_pipeline()], which chains [run_fastqc()], [run_pear_merge()], [run_minimap2()], [summarise_bam_stats()], and optionally [run_multiqc()]. External tools are provisioned into a conda environment by [configure_environment()] and activated for the current R session by [use_barbac_env()].
Visualisation
[barbac_ts_area()] renders stacked-area plots of lineage trajectories over time. It accepts long-format or Bartender-wide input, supports late-appearing lineages via grid completion, and fills missing (barcode, timepoint) cells with either zero or a small epsilon.
References
Theodosiou L, Farr AD, Rainey PB (2026). *barbac: A versatile tool for analysing DNA barcode sequences.* Bioinformatics (in preparation).
Johnson MS, Venkataram S, Kryazhimskiy S (2023). *Best Practices in Designing, Sequencing, and Identifying Random DNA Barcodes.* J Mol Evol 91:263-280.
Tavakolian N, Cochran JR, Levy SF (2022). *Shepherd: accurate clustering for correcting DNA barcode errors.* Bioinformatics 38:3710-3716.
Author
Maintainer: Loukas Theodosiou theodosiou@evolbio.mpg.de
Examples
if (FALSE) { # \dontrun{
# Cluster a barcode count table
result <- super_cluster2("reads.csv", distance = 3, merge_ratio = 20)
# Visualise a time series of clustered barcode frequencies
barbac_ts_area(time_series_df, min_total_count = 10)
} # }