scran::ComputeMedianMad Class Reference

Compute the median and MAD from an array of values. More...

#include <ComputeMedianMad.hpp>

Classes
struct	Defaults
	Default parameters. More...
struct	Results
	Medians and MADs, possibly for multiple blocks. More...

Public Member Functions
ComputeMedianMad &	set_log (bool l=Defaults::log)
ComputeMedianMad &	set_median_only (bool m=Defaults::median_only)
template<typename Input , typename Buffer >
Results	run (size_t n, const Input *metrics, Buffer *buffer) const
template<typename Input >
Results	run (size_t n, const Input *metrics) const
template<typename Input , typename Block , typename Buffer >
Results	run_blocked (size_t n, const Block *block, std::vector< int > starts, const Input *metrics, Buffer *buffer) const
template<typename Block , typename Input >
Results	run_blocked (size_t n, const Block *block, const Input *metrics) const

Static Public Member Functions
template<typename Block >
static std::vector< int >	compute_block_starts (size_t n, const Block *block)

Detailed Description

Compute the median and MAD from an array of values.

Pretty much as it says on the can; computes the median of an array of values first, and uses the median to then compute the median absolute deviation (MAD) from that array. We support calculation of medians and MADs for separate "blocks" of observations in the same array. These statistics can be used in FilterOutliers to obtain filter thresholds for removing outliers.

Member Function Documentation

set_log()

ComputeMedianMad & scran::ComputeMedianMad::set_log ( bool  l = Defaults::log )

inline

Parameters

l	Whether to compute the median and MAD after log-transformation of the values. This is useful for defining thresholds based on fold changes from the center. If true, all values are assumed to be non-negative.

Returns

A reference to this ComputeMedianMad object.

set_median_only()

ComputeMedianMad & scran::ComputeMedianMad::set_median_only ( bool  m = Defaults::median_only )

inline

Parameters

m	Whether to only compute the median. If true, Results::mads will be filled with NaNs.

Returns

A reference to this ComputeMedianMad object.

compute_block_starts()

template<typename Block >

static std::vector< int > scran::ComputeMedianMad::compute_block_starts ( size_t  n, const Block *  block  )

inlinestatic

Template Parameters

Block

Integer type, containing the block IDs.

Parameters

	n	Number of observations.
[in]	block	Pointer to an array of block identifiers. If provided, the array should be of length equal to n. Values should be integer IDs in where is the number of blocks. If a null pointer is supplied, all observations are assumed to belong to the same block.

Returns

Vector of indices specifying the start position for each block on the workspace buffer, if observations from successive blocks were arranged contiguously onto the buffer.

run() [1/2]

template<typename Input , typename Buffer >

Results scran::ComputeMedianMad::run ( size_t  n, const Input *  metrics, Buffer *  buffer  ) const

inline

Template Parameters

Input	Numeric type for the input.
Buffer	Floating-point type for the buffer.

Parameters

	n	Number of observations.
[in]	metrics	Pointer to an array of observations of length n. NaNs are ignored.
	buffer	Pointer to an array of length n, to be used as a workspace.

Returns

Median and MAD for the metrics.

run() [2/2]

template<typename Input >

Results scran::ComputeMedianMad::run ( size_t  n, const Input *  metrics  ) const

inline

This is an overloaded member function, provided for convenience. It differs from the above function only in what argument(s) it accepts.

Template Parameters

Input

Numeric type for the input.

Parameters

	n	Number of observations.
[in]	metrics	Pointer to an array of observations of length n. NaNs are ignored.

Returns

Median and the MAD for metrics.

run_blocked() [1/2]

template<typename Input , typename Block , typename Buffer >

Results scran::ComputeMedianMad::run_blocked ( size_t  n, const Block *  block, std::vector< int >  starts, const Input *  metrics, Buffer *  buffer  ) const

inline

Template Parameters

Block	Integer type, containing the block IDs.
Input	Numeric type for the input.
Buffer	Floating-point type for the buffer.

Parameters

	n	Number of observations.
[in]	block	Optional pointer to an array of block identifiers. If provided, the array should be of length equal to ncells. Values should be integer IDs in where is the number of blocks. If a null pointer is supplied, all observations are assumed to belong to the same block.
	starts	Vector of start positions for the observations from each block, typically generated by calling compute_block_starts() on n and block.
[in]	metrics	Pointer to an array of observations of length n.
	buffer	Pointer to an array of length n, to be used as a workspace.

Returns

Medians and MADs for each block in metrics.

run_blocked() [2/2]

template<typename Block , typename Input >

Results scran::ComputeMedianMad::run_blocked ( size_t  n, const Block *  block, const Input *  metrics  ) const

inline

This is an overloaded member function, provided for convenience. It differs from the above function only in what argument(s) it accepts.

Template Parameters

Block	Integer type, containing the block IDs.
Input	Type of observation, numeric.

Parameters

	n	Number of observations.
[in]	block	Optional pointer to an array of block identifiers. If provided, the array should be of length equal to ncells. Values should be integer IDs in where is the number of blocks. If a null pointer is supplied, all observations are assumed to belong to the same block.
[in]	metrics	Pointer to an array of observations of length n.

Returns

Median and the MAD for each block in metrics.

---

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

• scran/quality_control/ComputeMedianMad.hpp