BigBacter is a bacterial genomic surveillance pipeline that can:
🧬 Iterative clustering - cluster assignments stay consistent across runs using a per-sample sourmash database that expands automatically with each new submission
🧬 Soft-core phylogenomics - retains substantially more phylogenetic signal than strict-core approaches by tolerating a configurable level of missing data
🧬 Automated reference selection - selects the most representative assembly per cluster using k-mer containment and assembly quality scoring; reuses the same reference on subsequent runs for consistent SNP distances
🧬 Dual distance metrics - reports both core-genome SNP distances and whole-genome containment scores to capture both SNP-level and accessory genome variation
BigBacter accepts reads, assemblies, or SRA and GenBank accessions, and fills in whatever is missing — assembling, subsampling, and assigning species-level taxonomy along the way. Because clustering is k-mer based and stored per-sample, no reference database has to be built before a new species can be run, and a run can start with as little as a single isolate. BigBacter is designed to run downstream of a workflow with robust assembly QC, such as PHoeNIx, TheiaProk, or Bactopia.
See the docs for more information.