Quick start

All three examples use the sample data in the repository’s data/ directory, and assume target/release/cuttlefish is on your PATH as cuttlefish.

An uncolored graph from references

mkdir -p work
cuttlefish build \
  --ref \
  --seq data/refs1.fa \
  --kmer-len 7 \
  --min-len 3 \
  --threads 16 \
  --work-dir work \
  --output graph

The maximal unitigs are written to graph.fa.

A colored graph from a collection

cuttlefish build \
  --ref \
  --list genomes.list \
  --kmer-len 31 \
  --threads 32 \
  --max-memory 128 \
  --work-dir work \
  --output colored-graph \
  --color

Each path in genomes.list becomes one source color. The build writes colored-graph.fa plus the color repository colored-graph.cf3rs.color-repository/ beside it. See Output formats for how the colors are encoded, and read them back with cuttlefish colors:

# unitigs carrying source 3 but not source 7
cuttlefish colors grep -r colored-graph.cf3rs.color-repository \
    -i colored-graph.fa --all-of 3 --none-of 7

A graph from sequencing reads

Select --read instead of --ref; the default (k + 1)-mer frequency cutoff then changes from 1 to 2, so singleton k-mers from sequencing error are dropped:

cuttlefish build \
  --read \
  --seq reads.fastq.gz \
  --cutoff 2 \
  --work-dir work \
  --output reads-graph

Checking two runs agree

Unitigs may be emitted in either orientation, and cyclic ones at any rotation, so diff is not a valid comparison. Use the compare subcommand:

cuttlefish compare -a graph-a.fa -b graph-b.fa --kmer-len 31 --work-dir cmp

See Comparing graphs.