LocARNA::MatchProbs Class Reference

Provides probabilities for each match. More...

#include <match_probs.hh>

List of all members.

Classes
class	ProbConsParameter
	Maintains parameter for computing match probabilities.
Public Types
typedef size_t	size_type
	size
Public Member Functions
	MatchProbs ()
	construct as empty object
	MatchProbs (const std::string &filename)
	construct from file
void	pairHMM_probs (const Sequence &seqA, const Sequence &seqB, const std::string &file)
void	pf_probs (const RnaData &rnaA, const RnaData &rnaB, const Matrix< double > &sim_mat, const Alphabet< char > &alphabet, double gap_opening, double gap_extension, double pf_struct_weight, double temp, bool flag_local)
std::istream &	read (std::istream &in)
void	read (const std::string &filename)
	read the probabilities from a file
std::istream &	read_sparse (std::istream &in, size_type lenA, size_type lenB)
void	read_sparse (const std::string &filename, size_type lenA, size_type lenB)
std::ostream &	write (std::ostream &out) const
void	write (const std::string &filename) const
	write the probabilities to a file
std::ostream &	write_sparse (std::ostream &out, double threshold) const
void	write_sparse (const std::string &filename, double threshold) const
size_type	get_lenA () const
	get the length of the first sequence
size_type	get_lenB () const
	get the length of the second sequence
double	prob (size_t i, size_t j) const
	return the match probability for the two bases

---

Detailed Description

Provides probabilities for each match.

The probabilities are either computed or read in from file

for computing probabilities, we offer two methods:

1) For computing the probabilities the class uses a pairHMM analogously to PROBCONS. Also, the class reads transition probabilities from a file in the format of Probcons, which allows to use their parameter files

2) similar to proba/probalign, use statistical-mechanics-like model. Assume alignments are Boltzman distributed, calc match probs via partition function

the second approach supports Stral-like scoring (using pf_struct_weight as "alpha")

---

Member Function Documentation

void LocARNA::MatchProbs::pairHMM_probs	(	const Sequence &	seqA,
		const Sequence &	seqB,
		const std::string &	file
	)

read probcons parameter file and compute match probabilities for the two given sequences

Exceptions:

failure

void LocARNA::MatchProbs::pf_probs	(	const RnaData &	rnaA,
		const RnaData &	rnaB,
		const Matrix< double > &	sim_mat,
		const Alphabet< char > &	alphabet,
		double	gap_opening,
		double	gap_extension,
		double	pf_struct_weight,
		double	temp,
		bool	flag_local
	)

calculate edge probabilities via statistical mechanics model (partition function) get match scores via matrix sim_mat. The method accepts the matrix sim_mat together with an alphabet. The alphabet object is necessary to translate sequence symbols to the indices in this matrix.

std::istream& LocARNA::MatchProbs::read

(

std::istream &

in

)

read the probabilities from a stream assumes matrix starting 0,0 whereas sequences start 1,1

std::istream & LocARNA::MatchProbs::read_sparse	(	std::istream &	in,
		size_type	lenA,
		size_type	lenB
	)

read the probabilities from a stream read 'sparse' format "i j p"

void LocARNA::MatchProbs::read_sparse	(	const std::string &	filename,
		size_type	lenA,
		size_type	lenB
	)

read the probabilities from a file read 'sparse' format "i j p"

std::ostream& LocARNA::MatchProbs::write

(

std::ostream &

out

)

const

write the probabilities to a stream writes matrix starting 0,0 whereas sequences start 1,1

std::ostream & LocARNA::MatchProbs::write_sparse	(	std::ostream &	out,
		double	threshold
	)		const

write the probabilities to a stream, only probs >= threshold use format "i j p"

void LocARNA::MatchProbs::write_sparse	(	const std::string &	filename,
		double	threshold
	)		const

write the probabilities to a file, only probs >= threshold use format "i j p"

---

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

• src/LocARNA/match_probs.hh
• src/LocARNA/match_probs.cc