Crate gecco 

GECCO: Gene Cluster prediction with Conditional Random Fields.

This crate provides both a CLI tool and a library API for identifying putative Biosynthetic Gene Clusters (BGCs) in genomic and metagenomic data.

Library usage

The easiest way to use GECCO as a library is through the Gecco struct:

use gecco::Gecco;
use gecco::orf::SeqRecord;

let pipeline = Gecco::builder()
    .data_dir("gecco_data")
    .threshold(0.8)
    .build()
    .unwrap();

let records = vec![SeqRecord {
    id: "contig_1".into(),
    seq: "ATGCCC...".into(),
}];

let results = pipeline.scan(&records).unwrap();
for cluster in &results.clusters {
    println!("{}: {} genes", cluster.id, cluster.genes.len());
}

For finer control, individual pipeline stages are also available as public methods on Gecco, or through the lower-level modules directly.

Re-exports

pub use model::Cluster;

pub use model::Domain;

pub use model::Gene;

pub use model::Protein;

pub use orf::SeqRecord;

pub use pipeline::Gecco;

pub use pipeline::GeccoBuilder;

pub use pipeline::GeccoResults;

Modules

cli

CLI commands for GECCO.

crf

CRF-based gene cluster prediction.

data_dir

Resolve the GECCO data directory containing HMM, CRF model, and InterPro files.

hmmer

HMMER domain annotation module.

interpro

InterPro metadata for domain annotations.

io

I/O utilities: TSV tables, compression, FASTA/GenBank.

model

Data layer types for gene cluster detection.

orf

ORF finding: gene prediction from DNA sequences.

output

pipeline

High-level pipeline API for library consumers.

refine

Cluster refinement: extract contiguous gene clusters from CRF predictions.

sklearn_rf

Evaluator and small trainer for scikit-learn-shaped random forests.

types

Supervised classifier to predict the type of a gene cluster.

util

Shared utility functions.