Package 'smartid' reference manual

Title:	Scoring and Marker Selection Method Based on Modified TF-IDF
Description:	This package enables automated selection of group specific signature, especially for rare population. The package is developed for generating specifc lists of signature genes based on Term Frequency-Inverse Document Frequency (TF-IDF) modified methods. It can also be used as a new gene-set scoring method or data transformation method. Multiple visualization functions are implemented in this package.
Authors:	Jinjin Chen [aut, cre]
Maintainer:	Jinjin Chen <[email protected]>
License:	MIT + file LICENSE
Version:	1.1.2
Built:	2024-07-26 03:14:36 UTC
Source:	https://github.com/bioc/smartid

calculate combined score

Description

compute TF (term/feature frequency), IDF (inverse document/cell frequency), IAE (inverse average expression of features) and combine the the final score

Usage

cal_score(
  data,
  tf = c("logtf", "tf"),
  idf = "prob",
  iae = "prob",
  slot = "counts",
  new.slot = "score",
  par.idf = NULL,
  par.iae = NULL
)

## S4 method for signature 'AnyMatrix'
cal_score(
  data,
  tf = c("logtf", "tf"),
  idf = "prob",
  iae = "prob",
  par.idf = NULL,
  par.iae = NULL
)

## S4 method for signature 'SummarizedExperiment'
cal_score(
  data,
  tf = c("logtf", "tf"),
  idf = "prob",
  iae = "prob",
  slot = "counts",
  new.slot = "score",
  par.idf = NULL,
  par.iae = NULL
)
cal_score(
  data,
  tf = c("logtf", "tf"),
  idf = "prob",
  iae = "prob",
  slot = "counts",
  new.slot = "score",
  par.idf = NULL,
  par.iae = NULL
)

## S4 method for signature 'AnyMatrix'
cal_score(
  data,
  tf = c("logtf", "tf"),
  idf = "prob",
  iae = "prob",
  par.idf = NULL,
  par.iae = NULL
)

## S4 method for signature 'SummarizedExperiment'
cal_score(
  data,
  tf = c("logtf", "tf"),
  idf = "prob",
  iae = "prob",
  slot = "counts",
  new.slot = "score",
  par.idf = NULL,
  par.iae = NULL
)

Arguments

`data`	an expression object, can be matrix or SummarizedExperiment
`tf`	a character, specify the TF method to use, can be "tf" or "logtf"
`idf`	a character, specify the IDF method to use. Available methods can be accessed using `idf_iae_methods()`
`iae`	a character, specify the IAE method to use. Available methods can be accessed using `idf_iae_methods()`
`slot`	a character, specify which slot to use when data is se object, optional, default 'counts'
`new.slot`	a character, specify the name of slot to save score in se object, optional, default 'score'
`par.idf`	other parameters for specified IDF methods
`par.iae`	other parameters for specified IAE methods

Value

A list of matrices or se object containing combined score

Examples

data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
cal_score(
  data,
  par.idf = list(label = sample(c("A", "B"), 10, replace = TRUE)),
  par.iae = list(label = sample(c("A", "B"), 10, replace = TRUE))
)
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
cal_score(
  data,
  par.idf = list(label = sample(c("A", "B"), 10, replace = TRUE)),
  par.iae = list(label = sample(c("A", "B"), 10, replace = TRUE))
)

Calculate score for each feature in each cell

Description

Calculate score for each feature in each cell

Usage

cal_score_init(
  expr,
  tf = c("logtf", "tf"),
  idf = "prob",
  iae = "prob",
  par.idf = NULL,
  par.iae = NULL
)
cal_score_init(
  expr,
  tf = c("logtf", "tf"),
  idf = "prob",
  iae = "prob",
  par.idf = NULL,
  par.iae = NULL
)

Arguments

`expr`	a count matrix, features in row and cells in column
`tf`	a character, specify the TF method to use, can be "tf" or "logtf"
`idf`	a character, specify the IDF method to use. Available methods can be accessed using `idf_iae_methods()`
`iae`	a character, specify the IAE method to use. Available methods can be accessed using `idf_iae_methods()`
`par.idf`	other parameters for specified IDF methods
`par.iae`	other parameters for specified IAE methods

Value

a list of combined score, tf, idf and iae

Examples

data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
label <- sample(c("A", "B"), 10, replace = TRUE)
smartid:::cal_score_init(data,
  par.idf = list(label = label),
  par.iae = list(label = label)
)
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
label <- sample(c("A", "B"), 10, replace = TRUE)
smartid:::cal_score_init(data,
  par.idf = list(label = label),
  par.iae = list(label = label)
)

compute overall score based on the given marker list

Description

compute overall score based on the given marker list

Usage

gs_score(data, features = NULL, slot = "score", suffix = "score")

## S4 method for signature 'AnyMatrix,ANY'
gs_score(data, features = NULL)

## S4 method for signature 'AnyMatrix,list'
gs_score(data, features = NULL, suffix = "score")

## S4 method for signature 'SummarizedExperiment,ANY'
gs_score(data, features = NULL, slot = "score", suffix = "score")
gs_score(data, features = NULL, slot = "score", suffix = "score")

## S4 method for signature 'AnyMatrix,ANY'
gs_score(data, features = NULL)

## S4 method for signature 'AnyMatrix,list'
gs_score(data, features = NULL, suffix = "score")

## S4 method for signature 'SummarizedExperiment,ANY'
gs_score(data, features = NULL, slot = "score", suffix = "score")

Arguments

`data`	an expression object, can be matrix or SummarizedExperiment
`features`	vector or named list, feature names to compute score
`slot`	a character, specify which slot to use when data is se object, optional, default 'score'
`suffix`	a character, specify the name suffix to save score when features is a named list

Value

A vector of overall score for each sample

Examples

data <- matrix(rnorm(100), 10, dimnames = list(seq_len(10)))
gs_score(data, features = seq_len(3))
data <- matrix(rnorm(100), 10, dimnames = list(seq_len(10)))
gs_score(data, features = seq_len(3))

Calculate scores of each cell on given features

Description

Calculate scores of each cell on given features

Usage

gs_score_init(score, features = NULL)
gs_score_init(score, features = NULL)

Arguments

`score`	matrix, features in row and samples in column
`features`	vector, feature names to compute score

Value

a vector of score

Examples

data <- matrix(rnorm(100), 10, dimnames = list(1:10))
gs_score_init(data, 1:5)
data <- matrix(rnorm(100), 10, dimnames = list(1:10))
gs_score_init(data, 1:5)

standard inverse average expression

Description

standard inverse average expression

Usage

iae(expr, features = NULL, thres = 0)
iae(expr, features = NULL, thres = 0)

Arguments

`expr`	a matrix, features in row and cells in column
`features`	vector, feature names or indexes to compute
`thres`	numeric, cell only counts when expr > threshold, default 0

Details

$\mathbf{IAE_i} = log(1+\frac{n}{\hat N_{i,j}+1})$

where $n$ is the total number of cells, $N_{i,j}$ is the counts of feature $i$ in cell $j$ .

Value

a vector of inverse average expression score for each feature

Examples

data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::iae(data)
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::iae(data)

inverse average expression using hdbscan cluster as label

Description

inverse average expression using hdbscan cluster as label

Usage

iae_hdb(expr, features = NULL, multi = TRUE, thres = 0, minPts = 2, ...)
iae_hdb(expr, features = NULL, multi = TRUE, thres = 0, minPts = 2, ...)

Arguments

`expr`	a matrix, features in row and cells in column
`features`	vector, feature names or indexes to compute
`multi`	logical, if to compute based on binary (FALSE) or multi-class (TRUE)
`thres`	numeric, cell only counts when expr > threshold, default 0
`minPts`	integer, minimum size of clusters, default 2. Details in `dbscan::hdbscan()`.
`...`	parameters for `dbscan::hdbscan()`

Details

Details as iae_prob().

Value

a matrix of inverse average expression score

Examples

set.seed(123)
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::iae_hdb(data)
set.seed(123)
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::iae_hdb(data)

labeled inverse average expression: IGM

Description

labeled inverse average expression: IGM

Usage

iae_igm(expr, features = NULL, label, lambda = 7, thres = 0)
iae_igm(expr, features = NULL, label, lambda = 7, thres = 0)

Arguments

`expr`	a matrix, features in row and cells in column
`features`	vector, feature names or indexes to compute
`label`	vector, group label of each cell
`lambda`	numeric, hyperparameter for IGM
`thres`	numeric, cell only counts when expr > threshold, default 0

Details

$\mathbf{IGM_i} = log(1+\lambda\frac{max(mean(N_{i,j\in D})_{k})}{\sum_{k}^{K}(mean(N_{i,j\in D})_{k}*r_{k})+e^{-8}})$

where $\lambda$ is the hyper parameter, $N_{i,j\in D}$ is the counts of feature $i$ in cell $j$ within class $D$ , and $r_k$ is the rank of $mean(N_{i,j\in D})$ .

Value

a vector of inverse gravity moment score for each feature

Examples

data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::iae_igm(data, label = sample(c("A", "B"), 10, replace = TRUE))
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::iae_igm(data, label = sample(c("A", "B"), 10, replace = TRUE))

inverse average expression: max

Description

inverse average expression: max

Usage

iae_m(expr, features = NULL, thres = 0)
iae_m(expr, features = NULL, thres = 0)

Arguments

`expr`	a matrix, features in row and cells in column
`features`	vector, feature names or indexes to compute
`thres`	numeric, cell only counts when expr > threshold, default 0

Details

$\mathbf{IAE_{i,j}} = log(1+\frac{max_{\{i^{'}\in j\}}(n_{i^{'}})}{\sum_{j = 1}^{n} max(0, N_{i,j} - threshold)+1})$

where $i$ is the feature $i$ and $i^{'}$ is the feature except $i$ , $N_{i,j}$ is the counts of feature $i$ in cell $j$ , and $n_{i^{'}}$ is $\sum_{j = 1}^{n} sign(N_{i,j} > threshold)$ .

Value

a matrix of inverse average expression score for each feature

Examples

data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::iae_m(data)
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::iae_m(data)

labeled inverse average expression: probability based

Description

labeled inverse average expression: probability based

Usage

iae_prob(expr, features = NULL, label, multi = TRUE, thres = 0)
iae_prob(expr, features = NULL, label, multi = TRUE, thres = 0)

Arguments

`expr`	a matrix, features in row and cells in column
`features`	vector, feature names or indexes to compute
`label`	vector, group label of each cell
`multi`	logical, if to compute based on binary (FALSE) or multi-class (TRUE)
`thres`	numeric, cell only counts when expr > threshold, default 0

Details

$\mathbf{IAE_{i,j}} = log(1+\frac{mean(N_{i,j\in D})}{max(mean(N_{i,j\in \hat D}))+ e^{-8}}*mean(N_{i,j\in D}))$

where $N_{i,j\in D}$ is the counts of feature $i$ in cell $j$ within class $D$ , and $\hat D$ is the class except $D$ .

Value

a matrix of inverse average expression score

Examples

data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::iae_prob(data, label = sample(c("A", "B"), 10, replace = TRUE))
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::iae_prob(data, label = sample(c("A", "B"), 10, replace = TRUE))

labeled inverse average expression: relative frequency

Description

labeled inverse average expression: relative frequency

Usage

iae_rf(expr, features = NULL, label, multi = TRUE, thres = 0)
iae_rf(expr, features = NULL, label, multi = TRUE, thres = 0)

Arguments

`expr`	a matrix, features in row and cells in column
`features`	vector, feature names or indexes to compute
`label`	vector, group label of each cell
`multi`	logical, if to compute based on binary (FALSE) or multi-class (TRUE)
`thres`	numeric, cell only counts when expr > threshold, default 0

Details

$\mathbf{IAE} = log(1+\frac{mean(N_{i,j\in D})}{max(mean(N_{i,j\in \hat D}))+ e^{-8}})$

where $N_{i,j\in D}$ is the counts of feature $i$ in cell $j$ within class $D$ , and $\hat D$ is the class except $D$ .

Value

a matrix of inverse average expression score

Examples

data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::iae_rf(data, label = sample(c("A", "B"), 10, replace = TRUE))
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::iae_rf(data, label = sample(c("A", "B"), 10, replace = TRUE))

inverse average expression using standard deviation (SD)

Description

inverse average expression using standard deviation (SD)

Usage

iae_sd(expr, features = NULL, log = FALSE, thres = 0)
iae_sd(expr, features = NULL, log = FALSE, thres = 0)

Arguments

`expr`	a matrix, features in row and cells in column
`features`	vector, feature names or indexes to compute
`log`	logical, if to do log-transformation
`thres`	numeric, cell only counts when expr > threshold, default 0

Details

$\mathbf{IAE} = log(1+sd(tf_{i})*\frac{n}{\sum_{j=1}^{n}max(0,N_{i,j})+1})$

where $tf_i$ is the term frequency of feature $i$ , see details in tf(), $n$ is the total number of cells and $N_{i,j}$ is the counts of feature $i$ in cell $j$ .

Value

a vector of inverse average expression score for each feature

Examples

data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::iae_sd(data)
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::iae_sd(data)

standard inverse cell frequency

Description

standard inverse cell frequency

Usage

idf(expr, features = NULL, thres = 0)
idf(expr, features = NULL, thres = 0)

Arguments

`expr`	a matrix, features in row and cells in column
`features`	vector, feature names or indexes to compute
`thres`	numeric, cell only counts when expr > threshold, default 0

Details

$\mathbf{IDF_i} = log(1+\frac{n}{n_i+1})$

where $n$ is the total number of cells, $n_i$ is the number of cells containing feature i.

Value

a vector of inverse cell frequency score for each feature

Examples

data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::idf(data)
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::idf(data)

inverse document frequency using hdbscan cluster as label

Description

inverse document frequency using hdbscan cluster as label

Usage

idf_hdb(expr, features = NULL, multi = TRUE, thres = 0, minPts = 2, ...)
idf_hdb(expr, features = NULL, multi = TRUE, thres = 0, minPts = 2, ...)

Arguments

`expr`	a matrix, features in row and cells in column
`features`	vector, feature names or indexes to compute
`multi`	logical, if to compute based on binary (FALSE) or multi-class (TRUE)
`thres`	numeric, cell only counts when expr > threshold, default 0
`minPts`	integer, minimum size of clusters, default 2. Details in `dbscan::hdbscan()`.
`...`	parameters for `dbscan::hdbscan()`

Details

Details as idf_prob().

Value

a matrix of inverse cell frequency score

Examples

set.seed(123)
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::idf_hdb(data)
set.seed(123)
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::idf_hdb(data)

Get names of available IDF and IAE methods

Description

Returns a named vector of IDF/IAE methods

Usage

idf_iae_methods()
idf_iae_methods()

Value

names of methods implemented

Examples

idf_iae_methods()
idf_iae_methods()

labeled inverse cell frequency: IGM

Description

labeled inverse cell frequency: IGM

Usage

idf_igm(expr, features = NULL, label, lambda = 7, thres = 0)
idf_igm(expr, features = NULL, label, lambda = 7, thres = 0)

Arguments

`expr`	a matrix, features in row and cells in column
`features`	vector, feature names or indexes to compute
`label`	vector, group label of each cell
`lambda`	numeric, hyperparameter for IGM
`thres`	numeric, cell only counts when expr > threshold, default 0

Details

$\mathbf{IGM_i} = log(1+\lambda\frac{max(n_{i,j\in D})_{k}}{\sum_{k}^{K}((n_{i,j\in D})_{k}*r_{k})+e^{-8}})$

where $\lambda$ is the hyper parameter, $n_{i,j\in D}$ is the number of cells containing feature $i$ in class $D$ , $r_k$ is the rank of $n_{i,j\in D}$ .

Value

a vector of inverse gravity moment score for each feature

Examples

data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::idf_igm(data, label = sample(c("A", "B"), 10, replace = TRUE))
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::idf_igm(data, label = sample(c("A", "B"), 10, replace = TRUE))

inverse cell frequency: max

Description

inverse cell frequency: max

Usage

idf_m(expr, features = NULL, thres = 0)
idf_m(expr, features = NULL, thres = 0)

Arguments

`expr`	a matrix, features in row and cells in column
`features`	vector, feature names or indexes to compute
`thres`	numeric, cell only counts when expr > threshold, default 0

Details

$\mathbf{IDF_{i,j}} = log(\frac{max_{\{i^{'}\in j\}}(n_{i^{'}})}{n_i+1})$

where $i$ is the feature $i$ and $i^{'}$ is the feature except $i$ , $n_i$ is the number of cells containing feature i, and $n_{i^{'}}$ is the number of cells containing feature $i^{'}$ .

Value

a matrix of inverse cell frequency score for each feature

Examples

data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::idf_m(data)
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::idf_m(data)

labeled inverse cell frequency: probability based

Description

labeled inverse cell frequency: probability based

Usage

idf_prob(expr, features = NULL, label, multi = TRUE, thres = 0)
idf_prob(expr, features = NULL, label, multi = TRUE, thres = 0)

Arguments

`expr`	a matrix, features in row and cells in column
`features`	vector, feature names or indexes to compute
`label`	vector, group label of each cell
`multi`	logical, if to compute based on binary (FALSE) or multi-class (TRUE)
`thres`	numeric, cell only counts when expr > threshold, default 0

Details

$\mathbf{IDF_{i,j}} = log(1+\frac{\frac{n_{i,j\in D}}{n_{j\in D}}}{max(\frac{n_{i,j\in \hat D}}{n_{j\in \hat D}})+ e^{-8}}\frac{n_{i,j\in D}}{n_{j\in D}})$

where $n_{i,j\in D}$ is the number of cells containing feature $i$ in class $D$ , $n_{j\in D}$ is the total number of cells in class $D$ , $\hat D$ is the class except $D$ .

Value

a matrix of inverse cell frequency score

Examples

data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::idf_prob(data, label = sample(c("A", "B"), 10, replace = TRUE))
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::idf_prob(data, label = sample(c("A", "B"), 10, replace = TRUE))

labeled inverse cell frequency: relative frequency

Description

labeled inverse cell frequency: relative frequency

Usage

idf_rf(expr, features = NULL, label, multi = TRUE, thres = 0)
idf_rf(expr, features = NULL, label, multi = TRUE, thres = 0)

Arguments

`expr`	a matrix, features in row and cells in column
`features`	vector, feature names or indexes to compute
`label`	vector, group label of each cell
`multi`	logical, if to compute based on binary (FALSE) or multi-class (TRUE)
`thres`	numeric, cell only counts when expr > threshold, default 0

Details

$\mathbf{IDF_{i,j}} = log(1+\frac{\frac{n_{i,j\in D}}{n_{j\in D}}}{max(\frac{n_{i,j\in \hat D}}{n_{j\in \hat D}})+ e^{-8}})$

where $n_{i,j\in D}$ is the number of cells containing feature $i$ in class $D$ , $n_{j\in D}$ is the total number of cells in class $D$ , $\hat D$ is the class except $D$ .

Value

a matrix of inverse cell frequency score

Examples

data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::idf_rf(data, label = sample(c("A", "B"), 10, replace = TRUE))
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::idf_rf(data, label = sample(c("A", "B"), 10, replace = TRUE))

inverse cell frequency using standard deviation (SD)

Description

inverse cell frequency using standard deviation (SD)

Usage

idf_sd(expr, features = NULL, log = FALSE, thres = 0)
idf_sd(expr, features = NULL, log = FALSE, thres = 0)

Arguments

`expr`	a matrix, features in row and cells in column
`features`	vector, feature names or indexes to compute
`log`	logical, if to do log-transformation
`thres`	numeric, cell only counts when expr > threshold, default 0

Details

$\mathbf{IDF_i} = log(1+sd(tf_{i})*\frac{n}{n_i+1})$

where $tf_i$ is the term frequency of feature $i$ , see details in tf(), $n$ is the total number of cells and $n_i$ is the number of cells containing feature $i$ .

Value

a vector of inverse cell frequency score for each feature

Examples

data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::idf_sd(data)
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::idf_sd(data)

select markers using HDBSCAN method

Description

select markers using HDBSCAN method

Usage

markers_hdbscan(
  top_markers,
  column = ".dot",
  s_thres = NULL,
  method = c("max.one", "remove.min"),
  minPts = 5,
  plot = FALSE,
  ...
)
markers_hdbscan(
  top_markers,
  column = ".dot",
  s_thres = NULL,
  method = c("max.one", "remove.min"),
  minPts = 5,
  plot = FALSE,
  ...
)

Arguments

`top_markers`	output of `top_markers()`
`column`	character, specify which column used as group label
`s_thres`	NULL or numeric, only features with score > threshold will be returned, if NULL will use 2 * average probability as threshold
`method`	can be "max.one" or "remove.min", if to only keep features in 1st component or return features not in the last component
`minPts`	integer, minimum size of clusters for `dbscan::hdbscan()`
`plot`	logical, if to plot mixture density and hist
`...`	other params for `dbscan::hdbscan()`

Value

a list of markers for each group

Examples

data <- matrix(rnorm(100), 10, dimnames = list(1:10))
top_n <- top_markers(data, label = rep(c("A", "B"), 5))
markers_hdbscan(top_n, minPts = 2)
data <- matrix(rnorm(100), 10, dimnames = list(1:10))
top_n <- top_markers(data, label = rep(c("A", "B"), 5))
markers_hdbscan(top_n, minPts = 2)

select markers using mclust EM method

Description

select markers using mclust EM method

Usage

markers_mclust(
  top_markers,
  column = ".dot",
  prob = 0.99,
  s_thres = NULL,
  method = c("max.one", "remove.min"),
  plot = FALSE,
  ...
)
markers_mclust(
  top_markers,
  column = ".dot",
  prob = 0.99,
  s_thres = NULL,
  method = c("max.one", "remove.min"),
  plot = FALSE,
  ...
)

Arguments

`top_markers`	output of `top_markers()`
`column`	character, specify which column used as group label
`prob`	numeric, probability cutoff for 1st component classification
`s_thres`	NULL or numeric, only features with score > threshold will be returned, if NULL will use 2 * average probability as threshold
`method`	can be "max.one" or "remove.min", if to only keep features in 1st component or return features not in the last component
`plot`	logical, if to plot mixture density and hist
`...`	other params for `mclust::densityMclust()`

Value

a list of markers for each group

Examples

data <- matrix(rnorm(100), 10, dimnames = list(1:10))
top_n <- top_markers(data, label = rep(c("A", "B"), 5))
markers_mclust(top_n)
data <- matrix(rnorm(100), 10, dimnames = list(1:10))
top_n <- top_markers(data, label = rep(c("A", "B"), 5))
markers_mclust(top_n)

select markers using mixtools EM method

Description

select markers using mixtools EM method

Usage

markers_mixmdl(
  top_markers,
  column = ".dot",
  prob = 0.99,
  k = 3,
  ratio = 2,
  dist = c("norm", "gamma"),
  maxit = 1e+05,
  plot = FALSE,
  ...
)
markers_mixmdl(
  top_markers,
  column = ".dot",
  prob = 0.99,
  k = 3,
  ratio = 2,
  dist = c("norm", "gamma"),
  maxit = 1e+05,
  plot = FALSE,
  ...
)

Arguments

`top_markers`	output of `top_markers()`
`column`	character, specify which column used as group label
`prob`	numeric, probability cutoff for 1st component classification
`k`	integer, number of components of mixtures
`ratio`	numeric, ratio cutoff of 1st component mu to 2nd component mu, only when ratio > cutoff will return markers for the group
`dist`	can be one of "norm" and "gamma", specify if to use `mixtools::normalmixEM()` or `mixtools::gammamixEM()`
`maxit`	integer, maximum number of iterations for EM
`plot`	logical, if to plot mixture density and hist
`...`	other params for `mixtools::normalmixEM()` or `mixtools::gammamixEM()`

Value

a list of markers for each group

Examples

set.seed(1000)
data <- matrix(rnorm(100), 10, dimnames = list(1:10))
top_n <- top_markers(data, label = rep(c("A", "B"), 5))
markers_mixmdl(top_n, k = 3)
set.seed(1000)
data <- matrix(rnorm(100), 10, dimnames = list(1:10))
top_n <- top_markers(data, label = rep(c("A", "B"), 5))
markers_mixmdl(top_n, k = 3)

boxplot of features overall score

Description

boxplot of features overall score

Usage

ova_score_boxplot(data, features, ref.group, label, method = "t.test")
ova_score_boxplot(data, features, ref.group, label, method = "t.test")

Arguments

`data`	matrix, features in row and samples in column
`features`	vector, feature names to plot
`ref.group`	character, reference group name
`label`	vector, group labels
`method`	character, statistical test to use, details in `ggpubr::stat_compare_means()`

Value

ggplot object

Examples

data <- matrix(rnorm(100), 10, dimnames = list(1:10))
ova_score_boxplot(data, 1:5, ref.group = "A", label = rep(c("A", "B"), 5))
data <- matrix(rnorm(100), 10, dimnames = list(1:10))
ova_score_boxplot(data, 1:5, ref.group = "A", label = rep(c("A", "B"), 5))

scale by mean of group mean for imbalanced data

Description

scale by mean of group mean for imbalanced data

Usage

scale_mgm(expr, label, pooled.sd = FALSE)
scale_mgm(expr, label, pooled.sd = FALSE)

Arguments

`expr`	matrix
`label`	a vector of group label
`pooled.sd`	logical, if to use pooled SD for scaling

Details

$z=\frac{x-\frac{\sum_k^{n_D}(\mu_k)}{n_D}}{s}$

where $\mu_k$ is the mean of x in $k^{th}$ class, and $n_D$ is the number of classes, $s$ is the standard deviation of x, when pooled.sd is set to be TRUE, $s$ will be replaced with $s_{pooled}$ , $s_{pooled}=\sqrt{\frac{\sum_k^{n_D}{(n_k-1){s_k}^2}}{\sum_k^{n_D}{n_k}-k}}$

Value

scaled matrix

Examples

scale_mgm(matrix(rnorm(100), 10), label = rep(letters[1:2], 5))
scale_mgm(matrix(rnorm(100), 10), label = rep(letters[1:2], 5))

barplot of processed score

Description

barplot of processed score

Usage

score_barplot(top_markers, column = ".dot", f_list, n = 30)
score_barplot(top_markers, column = ".dot", f_list, n = 30)

Arguments

`top_markers`	output of `top_markers()`
`column`	character, specify which column used as group label
`f_list`	a named list of markers
`n`	numeric, number of returned top genes for each group

Value

ggplot object

Examples

data <- matrix(rnorm(100), 10, dimnames = list(1:10))
top_n <- top_markers(data, label = rep(c("A", "B"), 5))
score_barplot(top_n)
data <- matrix(rnorm(100), 10, dimnames = list(1:10))
top_n <- top_markers(data, label = rep(c("A", "B"), 5))
score_barplot(top_n)

scRNA-seq test data of 4 groups simulated by `splatter`.

Description

A SingleCellExperiment object containing 4 groups with each group up-regulated DEGs saved in metadata.

Usage

data(sim_sce_test)
data(sim_sce_test)

Format

A SingleCellExperiment object of 100genes * 400 cells.

Value

SingleCellExperiment

Source

splatter::splatSimulate()

boxplot of split single feature score

Description

boxplot of split single feature score

Usage

sin_score_boxplot(data, features = NULL, ref.group, label, method = "t.test")
sin_score_boxplot(data, features = NULL, ref.group, label, method = "t.test")

Arguments

`data`	matrix, features in row and samples in column
`features`	vector, feature names to plot
`ref.group`	character, reference group name
`label`	vector, group labels
`method`	character, statistical test to use, details in `ggpubr::stat_compare_means()`

Value

faceted ggplot object

Examples

data <- matrix(rnorm(100), 10, dimnames = list(1:10))
sin_score_boxplot(data, 1:2, ref.group = "A", label = rep(c("A", "B"), 5))
data <- matrix(rnorm(100), 10, dimnames = list(1:10))
sin_score_boxplot(data, 1:2, ref.group = "A", label = rep(c("A", "B"), 5))

compute term/feature frequency within each cell

Description

compute term/feature frequency within each cell

Usage

tf(expr, log = FALSE)
tf(expr, log = FALSE)

Arguments

`expr`	a count matrix, features in row and cells in column
`log`	logical, if to do log-transformation

Details

$\mathbf{TF_{i,j}}=\frac{N_{i,j}}{\sum_j{N_{i,j}}}$

where $N_{i,j}$ is the counts of feature i in cell j.

Value

a matrix of term/gene frequency

Examples

data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::tf(data)
data <- matrix(rpois(100, 2), 10, dimnames = list(1:10))
smartid:::tf(data)

scale score and return top markers

Description

scale and transform score and output top markers for groups

Usage

top_markers(
  data,
  label,
  n = 10,
  use.glm = TRUE,
  scale = TRUE,
  use.mgm = TRUE,
  softmax = TRUE,
  slot = "score",
  ...
)

## S4 method for signature 'AnyMatrix'
top_markers(
  data,
  label,
  n = 10,
  use.glm = TRUE,
  scale = TRUE,
  use.mgm = TRUE,
  softmax = TRUE,
  slot = "score",
  ...
)

## S4 method for signature 'SummarizedExperiment'
top_markers(
  data,
  label,
  n = 10,
  use.glm = TRUE,
  scale = TRUE,
  use.mgm = TRUE,
  softmax = TRUE,
  slot = "score",
  ...
)
top_markers(
  data,
  label,
  n = 10,
  use.glm = TRUE,
  scale = TRUE,
  use.mgm = TRUE,
  softmax = TRUE,
  slot = "score",
  ...
)

## S4 method for signature 'AnyMatrix'
top_markers(
  data,
  label,
  n = 10,
  use.glm = TRUE,
  scale = TRUE,
  use.mgm = TRUE,
  softmax = TRUE,
  slot = "score",
  ...
)

## S4 method for signature 'SummarizedExperiment'
top_markers(
  data,
  label,
  n = 10,
  use.glm = TRUE,
  scale = TRUE,
  use.mgm = TRUE,
  softmax = TRUE,
  slot = "score",
  ...
)

Arguments

`data`	an expression object, can be matrix or SummarizedExperiment
`label`	a vector of group label
`n`	integer, number of returned top genes for each group
`use.glm`	logical, if to use `stats::glm()` to compute group mean score, if TRUE, also compute mean score difference as output
`scale`	logical, if to scale data by row
`use.mgm`	logical, if to scale data using `scale_mgm()`
`softmax`	logical, if to apply softmax transformation on output
`slot`	a character, specify which slot to use when data is se object, optional, default 'score'
`...`	params for `top_markers_abs()` or `top_markers_glm()`

Value

A tibble with top n feature names, group labels and ordered scores

Examples

data <- matrix(rgamma(100, 2), 10, dimnames = list(1:10))
top_markers(data, label = rep(c("A", "B"), 5))
data <- matrix(rgamma(100, 2), 10, dimnames = list(1:10))
top_markers(data, label = rep(c("A", "B"), 5))

calculate group median, MAD or mean score and order genes based on scores

Description

calculate group median, MAD or mean score and order genes based on scores

Usage

top_markers_abs(
  data,
  label,
  n = 10,
  pooled.sd = FALSE,
  method = c("median", "mad", "mean"),
  scale = TRUE,
  use.mgm = TRUE,
  softmax = TRUE,
  tau = 1
)
top_markers_abs(
  data,
  label,
  n = 10,
  pooled.sd = FALSE,
  method = c("median", "mad", "mean"),
  scale = TRUE,
  use.mgm = TRUE,
  softmax = TRUE,
  tau = 1
)

Arguments

`data`	matrix, features in row and samples in column
`label`	a vector of group label
`n`	integer, number of returned top genes for each group
`pooled.sd`	logical, if to use pooled SD for scaling
`method`	character, specify metric to compute, can be one of "median", "mad", "mean"
`scale`	logical, if to scale data by row
`use.mgm`	logical, if to scale data using `scale_mgm()`
`softmax`	logical, if to apply softmax transformation on output
`tau`	numeric, hyper parameter for softmax

Value

a tibble with feature names, group labels and ordered processed scores

Examples

data <- matrix(rgamma(100, 2), 10, dimnames = list(1:10))
top_markers_abs(data, label = rep(c("A", "B"), 5))
data <- matrix(rgamma(100, 2), 10, dimnames = list(1:10))
top_markers_abs(data, label = rep(c("A", "B"), 5))

calculate group mean score using glm and order genes based on scores difference

Description

calculate group mean score using glm and order genes based on scores difference

Usage

top_markers_glm(
  data,
  label,
  n = 10,
  family = gaussian(),
  scale = TRUE,
  use.mgm = TRUE,
  pooled.sd = FALSE,
  softmax = TRUE,
  tau = 1
)
top_markers_glm(
  data,
  label,
  n = 10,
  family = gaussian(),
  scale = TRUE,
  use.mgm = TRUE,
  pooled.sd = FALSE,
  softmax = TRUE,
  tau = 1
)

Arguments

`data`	matrix, features in row and samples in column
`label`	a vector of group label
`n`	integer, number of returned top genes for each group
`family`	family for glm, details in `stats::glm()`
`scale`	logical, if to scale data by row
`use.mgm`	logical, if to scale data using `scale_mgm()`
`pooled.sd`	logical, if to use pooled SD for scaling
`softmax`	logical, if to apply softmax transformation on output
`tau`	numeric, hyper parameter for softmax

Value

a tibble with feature names, group labels and ordered processed scores

Examples

data <- matrix(rgamma(100, 2), 10, dimnames = list(1:10))
top_markers_glm(data, label = rep(c("A", "B"), 5))
data <- matrix(rgamma(100, 2), 10, dimnames = list(1:10))
top_markers_glm(data, label = rep(c("A", "B"), 5))

compute group summarized score and order genes based on processed scores

Description

compute group summarized score and order genes based on processed scores

Usage

top_markers_init(
  data,
  label,
  n = 10,
  use.glm = TRUE,
  scale = TRUE,
  use.mgm = TRUE,
  softmax = TRUE,
  ...
)
top_markers_init(
  data,
  label,
  n = 10,
  use.glm = TRUE,
  scale = TRUE,
  use.mgm = TRUE,
  softmax = TRUE,
  ...
)

Arguments

`data`	matrix, features in row and samples in column
`label`	a vector of group label
`n`	integer, number of returned top genes for each group
`use.glm`	logical, if to use `stats::glm()` to compute group mean score, if TRUE, also compute mean score difference as output
`scale`	logical, if to scale data by row
`use.mgm`	logical, if to scale data using `scale_mgm()`
`softmax`	logical, if to apply softmax transformation on output
`...`	params for `top_markers_abs()` or `top_markers_glm()`

Value

a tibble with feature names, group labels and ordered processed scores

Examples

data <- matrix(rgamma(100, 2), 10, dimnames = list(1:10))
top_markers_init(data, label = rep(c("A", "B"), 5))
data <- matrix(rgamma(100, 2), 10, dimnames = list(1:10))
top_markers_init(data, label = rep(c("A", "B"), 5))

Package 'smartid'

Help Index

calculate combined score

Description

Usage

Arguments

Value

Examples

Calculate score for each feature in each cell

Description

Usage

Arguments

Value

Examples

compute overall score based on the given marker list

Description

Usage

Arguments

Value

Examples

Calculate scores of each cell on given features

Description

Usage

Arguments

Value

Examples

standard inverse average expression

Description

Usage

Arguments

Details

Value

Examples

inverse average expression using hdbscan cluster as label

Description

Usage

Arguments

Details

Value

Examples

labeled inverse average expression: IGM

Description

Usage

Arguments

Details

Value

Examples

inverse average expression: max

Description

Usage

Arguments

Details

Value

Examples

labeled inverse average expression: probability based

Description

Usage

Arguments

Details

Value

Examples

labeled inverse average expression: relative frequency

Description

Usage

Arguments

Details

Value

Examples

inverse average expression using standard deviation (SD)

Description

Usage

Arguments

Details

Value

Examples

standard inverse cell frequency

Description

Usage

Arguments

Details