// the take
FALCON-Phase is a Snakemake pipeline and C core that combines Hi-C read density with PacBio FALCON-Unzip long-read assemblies to resolve haplotype-switch errors and produce phased, chromosome-scale diploid scaffolds. It's for computational biologists running genome assembly who already have a FALCON-Unzip output and Hi-C data and need to correct phase switches rather than build an assembly from scratch.
The method is backed by a preprint with real validation on an F1 bull genome, and the README documents expected heterozygosity ranges (tested at 0.7%, degrades above ~5% divergence) rather than overselling applicability. The phasing core (binary matrix + results.txt with per-block read support and simulation proportions) is a legitimate statistical approach, not a black box. Snakemake orchestration with a provided SGE cluster config means the pipeline is reproducible and can scale to real HPC workloads out of the box.
Last commit is from November 2020 and the authors state outright in the README that v2 moved into the pb-assembly bioconda package, meaning this repo is a dead end by its own maintainers' admission. It's hard-coupled to FALCON-Unzip's specific contig naming convention (000000F / 000000F_001), so it's useless without that exact upstream tool rather than being a general Hi-C phasing utility. Setup requires manually editing config.json and config.sh with absolute paths to half a dozen external binaries (bwa, mummer, bedtools, samtools) instead of being packaged as a single installable unit, and the Hi-C mapping step is explicitly called out as an overnight job on 24-48 cores for mammalian genomes.
We read the README, the directory tree and a key source file in October 2026. Nobody pays to be listed. How we judge →