SeqWeightTable Class Reference

The SeqWeightTable class keeps track of sequence-specific bias parameters. More...

#include <biascorrection.h>

List of all members.

Public Member Functions
	SeqWeightTable (size_t window_size, size_t order, double alpha)
	SeqWeightTable Constructor.
	SeqWeightTable (size_t window_size, size_t order, std::string param_file_name, std::string identifier)
	A second constructor that loads the distribution from a parameter file.
void	copy_observed (const SeqWeightTable &other)
	A member function that overwrites the "observed" parameters with those from another SeqWeightTable.
void	copy_expected (const SeqWeightTable &other)
	A member function that overwrites the "expected" parameters with those from another SeqWeightTable.
void	increment_expected (const Sequence &seq, double mass, const std::vector< double > &fl_cdf)
	A member function that increments the expected counts for a sliding window through the given target sequence by some mass.
void	normalize_expected ()
	A member function that normalizes the expected counts and fills in the lower-ordered marginals.
void	increment_observed (const Sequence &seq, size_t i, double mass)
	A member function that increments the observed counts for the given fragment sequence by some (logged) mass.
double	get_weight (const Sequence &seq, size_t i) const
	A member function that calculates the bias weight (logged) of a window.
void	append_output (std::ofstream &outfile) const
	A member function that appends the marginal and conditional probabilities for the foreground and background Markov models to the given file, formatted in tables for easy readability.

---

Detailed Description

The SeqWeightTable class keeps track of sequence-specific bias parameters.

Allows for the bias associated with a given sequence to be calculated, and for the bias parameters to be updated based on additional observations. All values are stored in log space.

Author:

Adam Roberts

Date:

2011 Artistic License 2.0

Definition at line 33 of file biascorrection.h.

---

Constructor & Destructor Documentation

SeqWeightTable::SeqWeightTable	(	size_t	window_size,
		size_t	order,
		double	alpha
	)

SeqWeightTable Constructor.

Parameters:

window_size	an unsigned integer specifying the size of the bias window surrounding fragment ends.
order	a size_t specifying the order to use for the Markov chains modelling the sequence.
alpha	a double specifying the strength of the uniform prior (logged pseudo-counts for each parameter).

Definition at line 31 of file biascorrection.cpp.

SeqWeightTable::SeqWeightTable	(	size_t	window_size,
		size_t	order,
		std::string	param_file_name,
		std::string	identifier
	)

A second constructor that loads the distribution from a parameter file.

Note that the values should not be modified after using this constructor.

Parameters:

window_size	an unsigned integer specifying the size of the bias window surrounding fragment ends. Must match file.
order	a size_t specifying the order to use for the Markov chains modelling the sequence. Must match file.
param_file_name	a string specifying the path to the parameter file.
identifier	a string specifying the header for these parameters in the file.

---

Member Function Documentation

void SeqWeightTable::append_output

(

std::ofstream &

outfile

)

const

A member function that appends the marginal and conditional probabilities for the foreground and background Markov models to the given file, formatted in tables for easy readability.

Parameters:

outfile

the file to append to.

void SeqWeightTable::copy_expected

(

const SeqWeightTable &

other

)

A member function that overwrites the "expected" parameters with those from another SeqWeightTable.

Parameters:

other

another SeqWeightTable from which to copy the parameters.

Definition at line 95 of file biascorrection.cpp.

void SeqWeightTable::copy_observed

(

const SeqWeightTable &

other

)

A member function that overwrites the "observed" parameters with those from another SeqWeightTable.

Parameters:

other

another SeqWeightTable from which to copy the parameters.

Definition at line 91 of file biascorrection.cpp.

double SeqWeightTable::get_weight	(	const Sequence &	seq,
		size_t	i
	)		const

A member function that calculates the bias weight (logged) of a window.

This is the ratio of the observed and expected weights given by the two Markov models.

Parameters:

seq	the target sequence.
i	the central point of the bias window, ie the fragment end.

Returns:

The bias weight for the window.

Definition at line 114 of file biascorrection.cpp.

void SeqWeightTable::increment_expected	(	const Sequence &	seq,
		double	mass,
		const std::vector< double > &	fl_cdf
	)

A member function that increments the expected counts for a sliding window through the given target sequence by some mass.

Parameters:

seq	the target sequence.
mass	the amount to increment by in the parameter table.
fl_cdf	the fragment length CDF.

Definition at line 99 of file biascorrection.cpp.

void SeqWeightTable::increment_observed	(	const Sequence &	seq,
		size_t	i,
		double	mass
	)

A member function that increments the observed counts for the given fragment sequence by some (logged) mass.

Parameters:

seq	the target sequence (possibly reverse complemented) to which the fragment end maps.
i	the index into the sequence at which to center the bias window, ie where the fragment starts/ends.
mass	the amount to increment by (logged)

Definition at line 108 of file biascorrection.cpp.

---

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

• src/biascorrection.h
• src/biascorrection.cpp