#include <graph_caller.hpp>

Inheritance diagram for vg::FlowCaller:

Public Member Functions
	FlowCaller (const PathPositionHandleGraph &graph, SupportBasedSnarlCaller &snarl_caller, SnarlManager &snarl_manager, const string &sample_name, TraversalFinder &traversal_finder, const vector< string > &ref_paths, const vector< size_t > &ref_path_offsets, const vector< int > &ref_path_ploidies, AlignmentEmitter *aln_emitter, bool traversals_only, bool gaf_output, size_t trav_padding, bool genotype_snarls, const pair< size_t, size_t > &allele_length_range)
	Original constructor for non-nested mode. More...

	FlowCaller (const PathPositionHandleGraph &graph, SupportBasedSnarlCaller &snarl_caller, SnarlManager &snarl_manager, const string &sample_name, TraversalFinder &traversal_finder, const vector< string > &ref_paths, const vector< size_t > &ref_path_offsets, const vector< int > &ref_path_ploidies, AlignmentEmitter *aln_emitter, bool traversals_only, bool gaf_output, size_t trav_padding, bool genotype_snarls, const pair< size_t, size_t > &allele_length_range, bool nested, double cluster_threshold, bool cluster_post_genotype, bool star_allele)
	Extended constructor for nested mode with clustering and star alleles. More...

virtual	~FlowCaller ()

virtual bool	call_snarl (const Snarl &snarl)
	Call a given snarl, and print the output to out_stream. More...

virtual string	vcf_header (const PathHandleGraph &graph, const vector< string > &contigs, const vector< size_t > &contig_length_overrides={}) const
	Write the vcf header (version and contigs and basic info) More...

Public Member Functions inherited from vg::GraphCaller
	GraphCaller (SnarlCaller &snarl_caller, SnarlManager &snarl_manager)

virtual	~GraphCaller ()

virtual void	call_top_level_snarls (const HandleGraph &graph, RecurseType recurse_type=RecurseOnFail)

virtual void	call_top_level_chains (const HandleGraph &graph, size_t max_edges, size_t max_trivial, RecurseType recurise_type=RecurseOnFail)

void	set_show_progress (bool show_progress)
	toggle progress messages More...

Public Member Functions inherited from vg::VCFOutputCaller
	VCFOutputCaller (const string &sample_name)

virtual	~VCFOutputCaller ()

bool	add_variant (vcflib::Variant &var) const

void	write_variants (ostream &out_stream, const SnarlManager *snarl_manager=nullptr)

void	vcf_fixup (vcflib::Variant &var) const
	Run vcffixup from vcflib. More...

void	set_translation (const unordered_map< nid_t, pair< string, size_t >> *translation)
	Add a translation map. More...

void	set_nested (bool nested)
	Assume writing nested snarls is enabled. More...

Public Member Functions inherited from vg::GAFOutputCaller
	GAFOutputCaller (AlignmentEmitter *emitter, const string &sample_name, const vector< string > &ref_paths, size_t trav_padding)
	The emitter object is created and owned by external forces. More...

virtual	~GAFOutputCaller ()

void	emit_gaf_traversals (const PathHandleGraph &graph, const string &snarl_name, const vector< SnarlTraversal > &travs, int64_t ref_trav_idx, const string &ref_path_name, int64_t ref_path_position, const TraversalSupportFinder *support_finder=nullptr)
	print the GAF traversals More...

void	emit_gaf_variant (const PathHandleGraph &graph, const string &snarl_name, const vector< SnarlTraversal > &travs, const vector< int > &genotype, int64_t ref_trav_idx, const string &ref_path_name, int64_t ref_path_position, const TraversalSupportFinder *support_finder=nullptr)
	print the GAF genotype More...

SnarlTraversal	pad_traversal (const PathHandleGraph &graph, const SnarlTraversal &trav) const
	pad a traversal with (first found) reference path, adding up to trav_padding to each side More...

Protected Member Functions
bool	call_snarl_internal (const Snarl &snarl, const string &parent_ref_path_name, pair< size_t, size_t > parent_ref_interval, const ChildTraversalSets *parent_child_trav_sets=nullptr)

TraversalSet	find_child_traversal_set (const SnarlTraversal &parent_trav, const Snarl &child) const

Protected Member Functions inherited from vg::GraphCaller
vector< Chain >	break_chain (const HandleGraph &graph, const Chain &chain, size_t max_edges, size_t max_trivial)
	Break up a chain into bits that we want to call using size heuristics. More...

Protected Member Functions inherited from vg::VCFOutputCaller
void	add_allele_path_to_info (const HandleGraph *graph, vcflib::Variant &v, int allele, const Traversal &trav, bool reversed, bool one_based) const
	add a traversal to the VCF info field in the format of a GFA W-line or GAF path More...

void	add_allele_path_to_info (vcflib::Variant &v, int allele, const SnarlTraversal &trav, bool reversed, bool one_based) const
	legacy version of above More...

string	trav_string (const HandleGraph &graph, const SnarlTraversal &trav) const
	convert a traversal into an allele string More...

bool	emit_variant (const PathPositionHandleGraph &graph, SnarlCaller &snarl_caller, const Snarl &snarl, const vector< SnarlTraversal > &called_traversals, const vector< int > &genotype, int ref_trav_idx, const unique_ptr< SnarlCaller::CallInfo > &call_info, const string &ref_path_name, int ref_offset, bool genotype_snarls, int ploidy, function< string(const vector< SnarlTraversal > &, const vector< int > &, int, int, int)> trav_to_string=nullptr)

tuple< int64_t, int64_t, bool, step_handle_t, step_handle_t >	get_ref_interval (const PathPositionHandleGraph &graph, const Snarl &snarl, const string &ref_path_name) const

pair< string, int64_t >	get_ref_position (const PathPositionHandleGraph &graph, const Snarl &snarl, const string &ref_path_name, int64_t ref_path_offset) const
	used for making gaf traversal names More...

void	flatten_common_allele_ends (vcflib::Variant &variant, bool backward, size_t len_override) const

string	print_snarl (const HandleGraph *grpah, const handle_t &snarl_start, const handle_t &snarl_end, bool in_brackets=false) const

string	print_snarl (const Snarl &snarl, bool in_brackets=false) const
	legacy version of above More...

string	print_flipped_snarl (const Snarl &snarl, bool in_brackets=false) const
	The same as above, but print the snarl as if its orientation has been flipped. More...

void	scan_snarl (const string &allele_string, function< void(const string &, Snarl &)> callback) const

void	update_nesting_info_tags (const SnarlManager *snarl_manager)

Protected Attributes
const PathPositionHandleGraph &	graph
	the graph More...

TraversalFinder &	traversal_finder
	the traversal finder More...

vector< string >	ref_paths
	keep track of the reference paths More...

unordered_set< string >	ref_path_set

map< string, size_t >	ref_offsets
	keep track of offsets in the reference paths More...

map< string, int >	ref_ploidies
	keep traco of the ploidies (todo: just one map for all path stuff!!) More...

size_t	max_snarl_edges = 10000
	until we support nested snarls, cap snarl size we attempt to process More...

AlignmentEmitter *	alignment_emitter

bool	traversals_only
	toggle whether to genotype or just output the traversals More...

bool	gaf_output
	toggle whether to output vcf or gaf More...

bool	genotype_snarls

pair< size_t, size_t >	allele_length_range

bool	nested = false
	— Nested mode members — More...

double	cluster_threshold = 1.0
	cluster traversals with Jaccard similarity >= this threshold More...

bool	cluster_post_genotype = false
	if true, cluster after genotyping (for output grouping); if false, cluster before genotyping More...

bool	star_allele = false
	use * alleles for spanning haplotypes that don't traverse nested sites More...

Protected Attributes inherited from vg::GraphCaller
SnarlCaller &	snarl_caller
	Our Genotyper. More...

SnarlManager &	snarl_manager
	Our snarls. More...

bool	show_progress
	Toggle progress messages. More...

Protected Attributes inherited from vg::VCFOutputCaller
vcflib::VariantCallFile	output_vcf
	output vcf More...

string	sample_name
	Sample name. More...

vector< vector< pair< pair< string, size_t >, string > > >	output_variants

size_t	max_uncalled_alleles = 5
	print up to this many uncalled alleles when doing ref-genotpes in -a mode More...

const unordered_map< nid_t, pair< string, size_t > > *	translation

bool	include_nested

Protected Attributes inherited from vg::GAFOutputCaller
AlignmentEmitter *	emitter

string	gaf_sample_name
	Sample name. More...

size_t	trav_padding = 0

unordered_set< string >	ref_paths
	Reference paths are used to pad out traversals. If there are none, then first path found is used. More...

Additional Inherited Members
Public Types inherited from vg::GraphCaller
enum	RecurseType { RecurseOnFail , RecurseAlways , RecurseNever }

Static Protected Attributes inherited from vg::VCFOutputCaller
static const int64_t	max_vcf_line_length = 2000000000

Detailed Description

FlowCaller : Uses any traversals finder (ex, FlowTraversalFinder) to find traversals, and calls those based on how much support they have. Should work on any graph but will not report cyclic traversals. Supports nested calling when enabled with the nested flag, recursively processing child snarls. Designed to replace LegacyCaller, as it should miss fewer obviously good traversals, and is not dependent on old protobuf-based structures.

Constructor & Destructor Documentation

◆ FlowCaller() [1/2]

vg::FlowCaller::FlowCaller	(	const PathPositionHandleGraph &	graph,
		SupportBasedSnarlCaller &	snarl_caller,
		SnarlManager &	snarl_manager,
		const string &	sample_name,
		TraversalFinder &	traversal_finder,
		const vector< string > &	ref_paths,
		const vector< size_t > &	ref_path_offsets,
		const vector< int > &	ref_path_ploidies,
		AlignmentEmitter *	aln_emitter,
		bool	traversals_only,
		bool	gaf_output,
		size_t	trav_padding,
		bool	genotype_snarls,
		const pair< size_t, size_t > &	allele_length_range
	)

Original constructor for non-nested mode.

◆ FlowCaller() [2/2]

vg::FlowCaller::FlowCaller	(	const PathPositionHandleGraph &	graph,
		SupportBasedSnarlCaller &	snarl_caller,
		SnarlManager &	snarl_manager,
		const string &	sample_name,
		TraversalFinder &	traversal_finder,
		const vector< string > &	ref_paths,
		const vector< size_t > &	ref_path_offsets,
		const vector< int > &	ref_path_ploidies,
		AlignmentEmitter *	aln_emitter,
		bool	traversals_only,
		bool	gaf_output,
		size_t	trav_padding,
		bool	genotype_snarls,
		const pair< size_t, size_t > &	allele_length_range,
		bool	nested,
		double	cluster_threshold,
		bool	cluster_post_genotype,
		bool	star_allele
	)

Extended constructor for nested mode with clustering and star alleles.

◆ ~FlowCaller()

vg::FlowCaller::~FlowCaller ( )

virtual

Member Function Documentation

◆ call_snarl()

bool vg::FlowCaller::call_snarl ( const Snarl & snarl )

virtual

Call a given snarl, and print the output to out_stream.

Implements vg::GraphCaller.

◆ call_snarl_internal()

bool vg::FlowCaller::call_snarl_internal	(	const Snarl &	snarl,
		const string &	parent_ref_path_name,
		pair< size_t, size_t >	parent_ref_interval,
		const ChildTraversalSets *	parent_child_trav_sets = `nullptr`
	)

protected

Internal implementation of call_snarl that accepts parent context for nested mode When nested=true, this recursively calls children after processing the current snarl

Parameters

parent_ref_path_name	Reference path from parent (for off-reference snarls)
parent_ref_interval	Reference interval from parent
parent_child_trav_sets	If non-null, contains one TraversalSet per parent allele. Each set contains all traversals through this child that are consistent with that parent allele. The child genotypes by picking the best pair (one from each set) based on read support.

◆ find_child_traversal_set()

TraversalSet vg::FlowCaller::find_child_traversal_set	(	const SnarlTraversal &	parent_trav,
		const Snarl &	child
	)		const

protected

Find all traversals through a child snarl that are consistent with a parent traversal. "Consistent" means the child's entry/exit points match what's in the parent traversal. Uses the traversal finder to enumerate all valid paths through the child.

Parameters

parent_trav	The parent traversal defining entry/exit constraints
child	The child snarl to find traversals through

Returns: Set of traversals through child, empty if parent doesn't traverse child

◆ vcf_header()

string vg::FlowCaller::vcf_header	(	const PathHandleGraph &	graph,
		const vector< string > &	contigs,
		const vector< size_t > &	contig_length_overrides = `{}`
	)		const

virtual

Write the vcf header (version and contigs and basic info)

Reimplemented from vg::VCFOutputCaller.

Member Data Documentation

◆ alignment_emitter

AlignmentEmitter* vg::FlowCaller::alignment_emitter

protected

alignment emitter. if not null, traversals will be output here and no genotyping will be done

◆ allele_length_range

pair<size_t, size_t> vg::FlowCaller::allele_length_range

protected

clamp calling to alleles of a given length range more specifically, a snarl is only called if 1) its largest allele is >= allele_length_range.first and 2) all alleles are < allele_length_range.second

◆ cluster_post_genotype

bool vg::FlowCaller::cluster_post_genotype = false

protected

if true, cluster after genotyping (for output grouping); if false, cluster before genotyping

◆ cluster_threshold

double vg::FlowCaller::cluster_threshold = 1.0

protected

cluster traversals with Jaccard similarity >= this threshold

◆ gaf_output

bool vg::FlowCaller::gaf_output

protected

toggle whether to output vcf or gaf

◆ genotype_snarls

bool vg::FlowCaller::genotype_snarls

protected

toggle whether to genotype every snarl (by default, uncalled snarls are skipped, and coordinates are flattened out to minimize variant size – this turns all that off)

◆ graph

const PathPositionHandleGraph& vg::FlowCaller::graph

protected

the graph

◆ max_snarl_edges

size_t vg::FlowCaller::max_snarl_edges = 10000

protected

until we support nested snarls, cap snarl size we attempt to process

◆ nested

bool vg::FlowCaller::nested = false

protected

— Nested mode members —

enable recursive calling of child snarls

◆ ref_offsets

map<string, size_t> vg::FlowCaller::ref_offsets

protected

keep track of offsets in the reference paths

◆ ref_path_set

unordered_set<string> vg::FlowCaller::ref_path_set

protected

◆ ref_paths

vector<string> vg::FlowCaller::ref_paths

protected

keep track of the reference paths

◆ ref_ploidies

map<string, int> vg::FlowCaller::ref_ploidies

protected

keep traco of the ploidies (todo: just one map for all path stuff!!)

◆ star_allele

bool vg::FlowCaller::star_allele = false

protected

use * alleles for spanning haplotypes that don't traverse nested sites

◆ traversal_finder

TraversalFinder& vg::FlowCaller::traversal_finder

protected

the traversal finder

◆ traversals_only

bool vg::FlowCaller::traversals_only

protected

toggle whether to genotype or just output the traversals

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

src/graph_caller.hpp
src/graph_caller.cpp

Public Member Functions

Protected Member Functions

Protected Attributes

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

◆ FlowCaller() [1/2]

◆ FlowCaller() [2/2]

◆ ~FlowCaller()

Member Function Documentation

◆ call_snarl()

◆ call_snarl_internal()

◆ find_child_traversal_set()

◆ vcf_header()

Member Data Documentation

◆ alignment_emitter

◆ allele_length_range

◆ cluster_post_genotype

◆ cluster_threshold

◆ gaf_output

◆ genotype_snarls

◆ graph

◆ max_snarl_edges

◆ nested

◆ ref_offsets

◆ ref_path_set

◆ ref_paths

◆ ref_ploidies

◆ star_allele

◆ traversal_finder

◆ traversals_only