Bevel is a sequence similarity search tool that uses a minimizer database. Minimizers are representative kmers across a genomic region. Two databases can be queried against each other, resulting in a list of positions where two or more sequences match.

The code headers index.h and search.h have minimal dependences and can easily be incorporated into other software. The snippet below shows how to use the database.

struct ns * tDB = db_init();
struct ns * qDB = db_init();

loadOrBuild(tDB, “fasta.fa|fq”);
loadOrBuild(qDB, “fasta.fa|fq”);

search(tDB, qDB);

The inspiration and much of the code comes from minimap. If you find this code useful please cite:

Li, Heng. "Minimap and miniasm: fast mapping and de novo assembly for noisy long sequences." 
    Bioinformatics (2016): btw152.

installing

cd bevel && make

running

To produce all minimizer matches across two sequences:

bin/bevel target.fa query.fa

options

    -d <FLAG> Write databases to files
    -w <INT>  Save every Nth minimizer [100]
    -k <INT>  Minimizer size (up to 32) [17]

1. -d - Write the database to a file decreasing subsequent queries.
2. -w - Save a minimizer every [w] kmers.
3. -k - Minimizer size.

output

The output is a five column TSV that is written to STDOUT:

query Seqid	target Seqid	query Start	target Start	# minimizers found in target	# minimizers found in query