vg::WFAExtender Class Reference

#include <gbwt_extender.hpp>

Classes
struct	ErrorModel

Public Member Functions
	WFAExtender ()
	Create an empty WFAExtender. More...
	WFAExtender (const gbwtgraph::GBWTGraph &graph, const Aligner &aligner, const ErrorModel &error_model=default_error_model)
WFAAlignment	connect (std::string sequence, pos_t from, pos_t to) const
WFAAlignment	suffix (const std::string &sequence, pos_t from) const
WFAAlignment	prefix (const std::string &sequence, pos_t to) const

Public Attributes
const gbwtgraph::GBWTGraph *	graph
ReadMasker	mask
const Aligner *	aligner
const ErrorModel *	error_model

Static Public Attributes
static const ErrorModel	default_error_model
	If not specified, we use this default error model. More...

Detailed Description

A class that supports haplotype-consistent seed extension in a GBWTGraph using the WFA algorithm:

Marco-Sola, Moure, Moreto, Espinosa: Fast gap-affine pairwise alignment using the wavefront algorithm. Bioinformatics, 2021.

The algorithm either tries to connect two seeds or extends a seed to the start/end of the read.

WFAExtender also needs an Aligner object for scoring the extension candidates. While VG wants to maximize a four-parameter alignment score, WFA minimizes a three-parameter score. We use the conversion between the parameters from:

Eizenga, Paten: Improving the time and space complexity of the WFA algorithm and generalizing its scoring. bioRxiv, 2022.

VG scores a gap of length n as gap_open + (n - 1) * gap_extend, while WFA papers use gap_open + n * gap_extend. Hence we use gap_open - gap_extend as the effective four-parameter gap open score inside the aligner.

NOTE: Most internal arithmetic operations use 32-bit integers.

Constructor & Destructor Documentation

WFAExtender() [1/2]

vg::WFAExtender::WFAExtender

(

)

Create an empty WFAExtender.

WFAExtender() [2/2]

vg::WFAExtender::WFAExtender	(	const gbwtgraph::GBWTGraph &	graph,
		const Aligner &	aligner,
		const ErrorModel &	error_model = default_error_model
	)

Create a WFAExtender using the given GBWTGraph and Aligner objects. If an error model is passed, use that instead of the default error model. All arguments must outlive the WFAExtender.

Member Function Documentation

connect()

WFAAlignment vg::WFAExtender::connect	(	std::string	sequence,
		pos_t	from,
		pos_t	to
	)		const

Align the sequence to a haplotype between the two graph positions.

The endpoints are assumed to be valid graph positions. In order for there to be an alignment, there must be a haplotype that includes the endpoints and connects them. However, the endpoints are not covered by the returned alignment.

The sequence that will be aligned is passed by value. All non-ACGT characters are masked with character X, which should not match any character in the graph.

Returns a failed alignment if there is no alignment with an acceptable score.

NOTE: The alignment is to a path after from and before to. If the points are identical, such a path can only exist if there is a cycle.

prefix()

WFAAlignment vg::WFAExtender::prefix	(	const std::string &	sequence,
		pos_t	to
	)		const

A special case of connect() for aligning the sequence to a haplotype ending at the given position. If there is no alignment for the entire sequence with an acceptable score, returns the highest-scoring partial alignment, which may be empty.

Applies the full-length bonus if the result begins with a match or mismatch. TODO: Use the full-length bonus to determine the optimal alignment.

NOTE: This creates a prefix of the full alignment by aligning a suffix of the sequence.

suffix()

WFAAlignment vg::WFAExtender::suffix	(	const std::string &	sequence,
		pos_t	from
	)		const

A special case of connect() for aligning the sequence to a haplotype starting at the given position. If there is no alignment for the entire sequence with an acceptable score, returns the highest-scoring partial alignment, which may be empty.

Applies the full-length bonus if the result ends with a match or mismatch. TODO: Use the full-length bonus to determine the optimal alignment.

NOTE: This creates a suffix of the full alignment by aligning a prefix of the sequence.

Member Data Documentation

aligner

const Aligner* vg::WFAExtender::aligner

default_error_model

const WFAExtender::ErrorModel vg::WFAExtender::default_error_model

static

Initial value:

{
    WFAExtender::ErrorModel::default_mismatches(),
    WFAExtender::ErrorModel::default_gaps(),
    WFAExtender::ErrorModel::default_gap_length(),
    WFAExtender::ErrorModel::default_distance()
}

If not specified, we use this default error model.

error_model

const ErrorModel* vg::WFAExtender::error_model

graph

const gbwtgraph::GBWTGraph* vg::WFAExtender::graph

mask

ReadMasker vg::WFAExtender::mask

---

The documentation for this class was generated from the following files:

• src/gbwt_extender.hpp
• src/gbwt_extender.cpp