scran_pca::BlockedPcaResults< EigenMatrix_, EigenVector_ > Struct Template Reference

Results of blocked_pca(). More...

#include <blocked_pca.hpp>

Public Attributes
EigenMatrix_	components
EigenVector_	variance_explained
EigenVector_::Scalar	total_variance = 0
EigenMatrix_	rotation
EigenMatrix_	center
std::optional< EigenVector_ >	scale
irlba::Metrics	metrics

Detailed Description

template<typename EigenMatrix_, typename EigenVector_>
struct scran_pca::BlockedPcaResults< EigenMatrix_, EigenVector_ >

Results of blocked_pca().

Template Parameters

EigenMatrix_	A floating-point column-major Eigen::Matrix class.
EigenVector_	A floating-point Eigen::Vector class.

Member Data Documentation

center

template<typename EigenMatrix_ , typename EigenVector_ >

EigenMatrix_ scran_pca::BlockedPcaResults< EigenMatrix_, EigenVector_ >::center

Centering matrix. Each row corresponds to a block and each column corresponds to a gene (i.e., row of the input matrix). Each entry contains the mean of a particular gene in the corresponding block. For empty blocks, the mean for all genes is set to zero.

components

template<typename EigenMatrix_ , typename EigenVector_ >

EigenMatrix_ scran_pca::BlockedPcaResults< EigenMatrix_, EigenVector_ >::components

Matrix of principal component scores. By default, each row corresponds to a PC while each column corresponds to a cell in the input matrix. If BlockedPcaOptions::transpose = false, rows are cells instead.

The number of PCs is the smaller of BlockedPcaOptions::number and min(NR, NC) - 1, where NR and NC are the number of rows and columns, respectively, of the input matrix.

metrics

template<typename EigenMatrix_ , typename EigenVector_ >

irlba::Metrics scran_pca::BlockedPcaResults< EigenMatrix_, EigenVector_ >::metrics

Metrics for IRLBA, including whether the algorithm converged and the number of iterations/multiplications required.

rotation

template<typename EigenMatrix_ , typename EigenVector_ >

EigenMatrix_ scran_pca::BlockedPcaResults< EigenMatrix_, EigenVector_ >::rotation

Rotation matrix. Each row corresponds to a gene (i.e., row of the input matrix) while each column corresponds to a PC. The number of PCs is as described for BlockedPcaResults::components.

scale

template<typename EigenMatrix_ , typename EigenVector_ >

std::optional<EigenVector_> scran_pca::BlockedPcaResults< EigenMatrix_, EigenVector_ >::scale

Scaling vector, only returned if BlockedPcaOptions::scale = true. Each entry corresponds to a gene (i.e., row of the input matrix) and contains the scaling factor used to divide that gene's values if BlockedPcaOptions::scale = true. This is usually the weighted sum of the per-block sample standard deviation of that gene. For genes with zero variance in all blocks, the scaling factor is set to 1 to avoid non-finite values upon scaling. For input matrices with fewer than 2 cells, the scaling factor is set to 1 for all genes.

total_variance

template<typename EigenMatrix_ , typename EigenVector_ >

EigenVector_::Scalar scran_pca::BlockedPcaResults< EigenMatrix_, EigenVector_ >::total_variance = 0

Total variance of the dataset (possibly after scaling, if BlockedPcaOptions::scale = true). This can be used to divide variance_explained to obtain the percentage of variance explained.

variance_explained

template<typename EigenMatrix_ , typename EigenVector_ >

EigenVector_ scran_pca::BlockedPcaResults< EigenMatrix_, EigenVector_ >::variance_explained

Variance explained by each PC. Each entry corresponds to a column in components and is in decreasing order. The number of PCs is as described for BlockedPcaResults::components.

---

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

• scran_pca/blocked_pca.hpp