BACoN

Bait, Assemble and Compare Nanopore reads. BACoN takes Nanopore reads from many samples and a reference sequence, typically an organelle genome (chloroplast, mitochondrion) or another region of interest, and for each sample:

1. baits the reads that match the reference (minimap2, or BBDuk k-mers);
2. filters them by length and quality, and caps the depth (Filtlong);
3. assembles them: a templated consensus of the reads aligned to the reference (samtools, the default), or a de novo assembly (Flye or myloasm);

then compares the samples: SNPs (each assembly read along the reference with minimap2, the default for templated assemblies; SKA2 split k-mers, the default for de novo ones; or a Parsnp core-genome alignment), a matrix of pairwise SNP distances, a VCF of the SNPs relative to the reference, and a tree (FastTree or IQ-TREE). Everything is summarized in an HTML report, and in MultiQC sections.

BACoN was designed for genome skimming: low-coverage sequencing of total DNA, where the organelle reads are abundant enough to assemble. Enriching the target (targeted sequencing, adaptive sampling) gives more depth and better assemblies.

Page	Contents
Installation	conda environment, pip, checking the installation
Usage	inputs, all options, resuming, performance
Methods	what each step does, choosing an assembler and a SNP method, limits
Example	a small simulated dataset with known SNPs, and its report
Tutorial	a complete analysis of 28 potato cultivars from public data
Outputs	every file and column
Validation	how the tools were chosen, with simulated data of known truth and real data
FAQ	troubleshooting
Development	tests, continuous integration, releases