,
Kelly Street [aut] | Title: | Labels normalized cells for CyTOF data and assigns probabilities for each label |
|---|---|
| Description: | cytofQC is a package for initial cleaning of CyTOF data. It uses a semi-supervised approach for labeling cells with their most likely data type (bead, doublet, debris, dead) and the probability that they belong to each label type. This package does not remove data from the dataset, but provides labels and information to aid the data user in cleaning their data. Our algorithm is able to distinguish between doublets and large cells. |
| Authors: | Jill Lundell [aut, cre] ,
Kelly Street [aut] |
| Maintainer: | Jill Lundell <[email protected]> |
| License: | Artistic-2.0 |
| Version: | 1.7.0 |
| Built: | 2024-10-30 05:40:25 UTC |
| Source: | https://github.com/bioc/cytofQC |
Labels observations in a CyTOF dataset as a cell, gdpZero (zero for at least one Gaussian parameter), bead, debris, doublet, or dead cell.
Data from an fcs file are read directly into a
SingleCellExperiment using the readCytof
function.
The data can be labeled with a single function, labelQC, which
can be customized. Labeling can also be done using a set of other functions
that first select a set of events that clearly look like the event type
being modeled and then use those events to train a statistical learning
model that can identify the event type. These functions are discussed and
demonstrated in the vignette.
A plotting function called cytofHist is included that makes
assessing the characteristic of the data and labeling easy.
The package also includes a function called cytofQCreport
that generates a report of the labeling and can generate a umap created with
the QC variables and colored by event label.
| Package: | cytofQC |
| Type: | Package |
| Version: | 0.99.3 |
| Date: | 2022-11-11 |
| License: | Artistic-2.0 |
Maintainer: Jill Lundell <[email protected]>
Authors: J. Lundell, K. Street
Returns histogram for grouped data
cytofHist(x, group, type = c("count", "density"), na.rm = FALSE, title = NULL)cytofHist(x, group, type = c("count", "density"), na.rm = FALSE, title = NULL)
x |
Numeric vector of values that will be plotted. |
group |
A vector that contains the grouping variable. It can be a numeric, factor, or character vector. |
type |
Either "count" or "density". The "count" selection keeps the groups on the same scale. The "density" option will over emphasize the group with the fewest observations. This is helpful when identifying where certain subgroups are relative to the majority of the data. |
na.rm |
TRUE if NAs should be removed prior to plotting. FALSE if they should remain. The NAs will be plotted as a separate group if they are not removed. |
title |
Optional title for the plot |
A ggplot2 histogram.
data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = 'Beads', viability = c('cisPt1','cisPt2')) sce <- labelQC(sce) cytofHist(scores(sce, 'bead'), label(sce))data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = 'Beads', viability = c('cisPt1','cisPt2')) sce <- labelQC(sce) cytofHist(scores(sce, 'bead'), label(sce))
Generate a cytofQC report
cytofQCreport(x, outDir, sampName, runUMAP = TRUE, ...)cytofQCreport(x, outDir, sampName, runUMAP = TRUE, ...)
x |
A SingleCellExperiment object generated by |
outDir |
The output directory (currently required). |
sampName |
Basename of the output HTML file (if not provided, same as
|
runUMAP |
Logical value indicating whether or not to include a UMAP plot in the report. This plot can be beneficial for diagnostic purposes, but is time-consuming to generate. |
... |
Additions arguments that may be passed to the function. |
If successful, returns TRUE silently and generates the
specified QC report.
data("raw_data", package = "CATALYST") x <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) x <- labelQC(x, n = 500) tmp <- tempdir() cytofQCreport(x, tmp, 'example')data("raw_data", package = "CATALYST") x <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) x <- labelQC(x, n = 500) tmp <- tempdir() cytofQCreport(x, tmp, 'example')
Returns the final label assignments for a parameter using a gradient boosting machine
gbmLabel( x, type = c("bead", "doublet", "debris", "dead"), loss = c("auc", "class"), n = 4000, standardize = TRUE )gbmLabel( x, type = c("bead", "doublet", "debris", "dead"), loss = c("auc", "class"), n = 4000, standardize = TRUE )
x |
A |
type |
Identifies the type of label that is being modeled. Must be 'bead', 'doublet', 'debris', or 'dead'. |
loss |
Specifies the type of loss used to tune the GBM. Can be either "auc" for the area under the curve or "class" for classification error. |
n |
number of observations in training dataset. |
standardize |
Indicates if the data should be standardized. Because the data are on different scales, it should be standardized for this analysis. |
gbmLabel uses a gradient boosting machine to compute the final labels
for the specified parameter type (bead, doublet, debris, or dead). This step
cannot be completed until the corresponding initialization function
(initialBead, initialDebris, initialDoublet, or
initialDead) is done on the SingleCellExperiment created by
readCytof.
The gbm is tuned using eztune and then predicted
values are computed for all of the events in x. If the predicted
probability for the label type is greater than 0.5, the label is changed to
the specified type. However, if an observation already has a label other
than 'cell' in the label variable, it will not be changed. The
predicted probabilities for all of the observations are stored in the
variable associated with that type in the probs object of x
for further analysis. Thus, it is possible to have a probability greater
than 0.5 for 'debris' but still have a label of 'bead' if an observation
was classified as a bead prior to classifying the debris.
An updated SingleCellExperiment is returned with the labels
for the parameter of interest (bead, doublet, debris, or dead) added to
the label object of the SingleCellExperiment and the
probabilities for the event type added to the probs object of the
SingleCellExperiment.
data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) sce <- initialBead(sce) sce <- gbmLabel(sce, type = "bead", loss = "auc") head(probs(sce)) table(label(sce))data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) sce <- initialBead(sce) sce <- gbmLabel(sce, type = "bead", loss = "auc") head(probs(sce)) table(label(sce))
Preliminary bead classification
initialBead(x)initialBead(x)
x |
A |
The beads are typically the first cell classification that is done. The
different event types are labeled iteratively so the labels
vector in the colData will contain all of the labels and
probabilities computed up to this point. Only events that
have a "cell" label can be assigned an initial event classification of
"bead". This function computes a score that assesses how much an event
looks like a bead and then fits a mixture model to assign each event
a class of 1 for a bead, -1 for an event that is not a bead, or 0
for undetermined or previously assigned to a different event type.
The score is recorded in the score object in the colData and
the initial classification is recorded in the initial part of
the colData.
Each bead channel should classify into two fairly clear groups where one is the beads and the other is non-beads. A histogram of the bead score should show a clear, small peak that represents the beads.
A SingleCellExperiment that contains the bead score and the
bead designation for each event. This information is stored in the
score and initial objects in the colData for the
SingleCellExperiment.
data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = 'Beads', viability = c('cisPt1','cisPt2')) sce <- initialBead(sce) head(scores(sce)) head(initial(sce))data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = 'Beads', viability = c('cisPt1','cisPt2')) sce <- initialBead(sce) head(scores(sce)) head(initial(sce))
Preliminary viability classification
initialDead(x, dna = FALSE, standardize = TRUE)initialDead(x, dna = FALSE, standardize = TRUE)
x |
A |
dna |
If TRUE, the DNA will be used to determine viability. |
standardize |
A value of TRUE will use compute the doublet score using standardized data. The raw data will be used to compute the score if FALSE. It is highly recommended that the data are standardized prior to computing the score because the variables are on different scales. |
The beads are typically the first cell classification that is done because their identification is straightforward. Debris is typically classified after the beads and doublets classified after the debris. After the doublets are classified, the permeability is assessed. The permeability can be used to determine which cells are alive and which are dead. Dead cells are often gated out prior to gating debris and doublets. However, cleaning with respect to permeability is complicated in that it does not always make sense to clean "dead", or permeable, cells out of the data. Our default method chooses to classify them last because once an event is labeled "dead", it will not be assigned a different label. However, a user may wish to label the permeable, or dead cells right after cleaning the beads so that the "dead" label takes precedence over debris and doublets. Note that "dead" is used in place of permeable for labeling in this package.
Different event types are labeled iteratively so the labels
vector in the colData will contain all of the labels and
probabilities computed up to this point. Only events that
have a "cell" label can be assigned an initial event classification of
"dead". This function computes a score that assesses how much an event
looks like a permeable cell and then fits a mixture model to assign
each event a class of 1 for permeable, -1 for an event that is not
permeable, or 0 for undetermined or previously assigned to a different
event type. The score is recorded in the score object in the
colData and the initial classification is recorded in the initial
part of the colData.
The viability measures should classify into two fairly clear groups
where one is permeable cells and the other is non-permeability. DNA is
also often higher for cells that are permeable. The primary measure
for determining permeability is sum of the viability measures, but
the method allows for DNA content to be used as well.
The function initialGuess is used to determine the groups. The
members of the group with the largest mean are classified as 'dead'
and the rest are classified as not dead.
A SingleCellExperiment that contains the permeability score
and the permeability designation for each event. This information is stored
in the score and initial objects in the colData for the
SingleCellExperiment.
data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = 'Beads', viability = c('cisPt1','cisPt2')) sce <- initialBead(sce) sce <- initialDebris(sce) sce <- initialDoublet(sce) sce <- initialDead(sce) head(scores(sce)) head(initial(sce))data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = 'Beads', viability = c('cisPt1','cisPt2')) sce <- initialBead(sce) sce <- initialDebris(sce) sce <- initialDoublet(sce) sce <- initialDead(sce) head(scores(sce)) head(initial(sce))
Preliminary debris classification
initialDebris(x, score = c(1, 2, 3), standardize = TRUE)initialDebris(x, score = c(1, 2, 3), standardize = TRUE)
x |
A |
score |
A value of 1, 2, or 3 that specifies the debris score that should be calculated. See details for information on the debris score. |
standardize |
A value of TRUE will compute the debris score using standardized data. The raw data will be used to compute the score if FALSE. It is highly recommended that the data are standardized prior to computing the score because the variables are on different scales. |
The beads are typically the first cell classification that is done because their identification is straightforward. Debris is typically classified after the beads. This is because classifying debris is more straightforward than doublets and labeling them before the doublets aids in doublet classification.
Different event types are labeled iteratively so the labels
vector in the colData will contain all of the labels and
probabilities computed up to this point. Only events that
have a "cell" label can be assigned an initial event classification of
"debris". This function computes a score that assesses how much an event
looks like debris and then fits a mixture model to assign each event
a class of 1 for debris, -1 for an event that is not debris, or 0
for undetermined or previously assigned to a different event type.
The score is recorded in the score object in the colData and
the initial classification is recorded in the initial part of
the colData.
Several options are available for computing the debris score. The following list shows the debris score calculations. Each one can be selected by its number on the following list:
1 - (DNA1 + DNA2 + Event_length - Center - Width + Offset)
Residual + Offset - 2(DNA1) - 2(DNA2) - Event_length - Center - 0.5(Width)
1 - (DNA1 + DNA2 + Event_length)
A SingleCellExperiment that contains the debris score and the
debris designation for each event. This information is stored in the
score and initial objects in the colData for the
SingleCellExperiment.
data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = 'Beads', viability = c('cisPt1','cisPt2')) sce <- initialBead(sce) sce <- initialDebris(sce) head(scores(sce)) head(initial(sce))data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = 'Beads', viability = c('cisPt1','cisPt2')) sce <- initialBead(sce) sce <- initialDebris(sce) head(scores(sce)) head(initial(sce))
Preliminary doublet classification
initialDoublet(x, score = c(1, 2, 3), standardize = TRUE)initialDoublet(x, score = c(1, 2, 3), standardize = TRUE)
x |
A |
score |
A value of 1, 2, or 3 that specifies the doublet score that should be calculated. See details for information on the doublet score. |
standardize |
A value of TRUE will use compute the doublet score using standardized data. The raw data will be used to compute the score if FALSE. It is highly recommended that the data are standardized prior to computing the score because the variables are on different scales. |
The beads are typically the first cell classification that is done because their identification is straightforward. Debris is typically classified after the beads. This is because classifying debris is more straightforward than doublets and labeling them before the doublets aids in doublet classification.
Different event types are labeled iteratively so the labels
vector in the colData will contain all of the labels and
probabilities computed up to this point. Only events that
have a "cell" label can be assigned an initial event classification of
"doublet". This function computes a score that assesses how much an event
looks like a doublet and then fits a mixture model to assign each event
a class of 1 for doublet, -1 for an event that is not a doublet, or 0
for undetermined or previously assigned to a different event type.
The score is recorded in the score object in the colData and
the initial classification is recorded in the initial part of
the colData.
Several options are available for computing the doublet score. The following list shows the doublet score calculations. Each one can be selected by its number on the following list:
DNA1 + DNA2 + Residual + Event_length - Offset - 0.5(Width)
DNA1 + DNA2 + Residual + Event_length - Offset - 0.5(Width) + abs(Center)
0.3 * (DNA1 + DNA2 + Event_length) + Residual + Center + (max(Offset))
A SingleCellExperiment that contains the doublet score and
the doublet designation for each event. This information is stored in the
score and initial objects in the colData for the
SingleCellExperiment.
data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = 'Beads', viability = c('cisPt1','cisPt2')) sce <- initialBead(sce) sce <- initialDebris(sce) sce <- initialDoublet(sce) head(scores(sce)) head(initial(sce))data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = 'Beads', viability = c('cisPt1','cisPt2')) sce <- initialBead(sce) sce <- initialDebris(sce) sce <- initialDoublet(sce) head(scores(sce)) head(initial(sce))
General preliminary classification.
initialGuess(x, middleGroup = c(0, -1, 1))initialGuess(x, middleGroup = c(0, -1, 1))
x |
The score (ie. debris score, doublet score, etc.) to be used for predicting each event's label (eg. "doublet" vs. "cell"). |
middleGroup |
numeric. When the optimal model (according to BIC) is the
3-component mixture model, this argument determines how to assign the
middle group. Possible values are |
A list with the following elements:
label A
vector of the same length as x providing the labels (-1 for
cells, 1 for non-cells, 0 for uncertain).
fit1
Summary of the 1-component (half Normal) model fit.
fit2
Summary of the 2-component (half Normal + Normal) model fit.
fit1 Summary of the 3-component (half Normal + 2 Normals)
model fit.
data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) sce <- initialDoublet(sce) fit <- initialGuess(scores(sce, "doublet"))data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) sce <- initialDoublet(sce) fit <- initialGuess(scores(sce, "doublet"))
Returns the final label assignments the specified parameters
labelQC( x, model = c("svm", "rf", "gbm"), type = c("all", "bead", "doublet", "debris", "dead"), nTrain = 4000, loss = c("auc", "class") )labelQC( x, model = c("svm", "rf", "gbm"), type = c("all", "bead", "doublet", "debris", "dead"), nTrain = 4000, loss = c("auc", "class") )
x |
A |
model |
Type of model to use to do the labeling. Options are "svm" for a support vector machine, "gbm" for a gradient boosting machine, or "rf" for a random forest. |
type |
Types of events to model. Options are "all", "bead", "doublet", "debris", and "dead". |
nTrain |
The (maximum) number of data points to use when training a model to predict event types. |
loss |
Specifies the type of loss used to tune the GBM. Can be either "auc" for the area under the curve or "class" for classification error. This argument is ignored if random forest is used as the model. |
labelQC uses a support vector machine, gradient boosting machine,
or a random forest to compute the final labels
for the specified parameter types (bead, doublet, debris, or dead).
The predicted probabilities for all of the observations are stored in
the variable associated with that type for further analysis. Thus, it
is possible to have a probability greater than 0.5 for 'debris' but still
have a label of 'bead' if an observation was classified as a bead prior to
classifying the debris.
A SingleCellExperiment data.frame is returned with the
labels for the parameters of listed in types (bead, doublet, debris,
or dead) added to the label variable and the probabilities for
each of the columns pertaining to the parameters listed in probs.
data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) sce <- labelQC(sce) table(label(sce))data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) sce <- labelQC(sce) table(label(sce))
Returns indices for data to be used to create the final classification model
modelData(x, type = c("bead", "doublet", "debris", "dead"), n = 4000)modelData(x, type = c("bead", "doublet", "debris", "dead"), n = 4000)
x |
A |
type |
Identifies the type of label that is being modeled. Must be 'bead', 'doublet', 'debris', or 'dead'. Note that if no type of label is specified 'bead' will be used. |
n |
number of indices to return. |
The indices that are returned by modelData are be used to
create a model that can be used to classify the observations with
regard to the parameter of interest (bead, doublet, debris, dead).
It is used as part of gbmLabel, rfLabel,
svmLable, and labelQC. The function modelData
uses the score and the function initialGuess to randomly select
a set of data points that we are confident are of the event type and
not of the selected event type that can be used to train the data. Only
points that are labeled as -1 and 1 are considered for the
training dataset. The selected dataset is balance with a fairly equal
number of points from each group.
An integer vector that contains the indices of the events that should be included in the creation of the final classification model for the event type of interest (bead, debris, doublet, dead).
data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) sce <- initialBead(sce) train <- modelData(sce, type = "bead", n = 4000)data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) sce <- initialBead(sce) train <- modelData(sce, type = "bead", n = 4000)
Plot preliminary classification from initialGuess
plotInitialGuess( x, IG = NULL, fit = NULL, type = c("both", "full", "truncated") )plotInitialGuess( x, IG = NULL, fit = NULL, type = c("both", "full", "truncated") )
x |
The score (ie. debris score, doublet score, etc.) to be used for predicting each event's label (eg. "doublet" vs. "cell"). |
IG |
If NULL, the function |
fit |
If left blank or NULL, the best fit as determined by BIC will be
plotted. Otherwise, a numeric value of 1, 2, or 3 can be be selected to
plot single normal fit, mixture of two normals, or the mixture of three
normals as fit by |
type |
Type of graph to be plotted. If 'truncated' is selected, only half of the first normal distribution will be plotted. If 'both' is selected, both the truncated and full plot will be plotted. |
A histogram that shows the score with the mixture of normal distributions overlayed.
data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) sce <- initialDoublet(sce) plotInitialGuess(scores(sce, "doublet"), type = "both") plotInitialGuess(scores(sce, "doublet"), type = "truncated")data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) sce <- initialDoublet(sce) plotInitialGuess(scores(sce, "doublet"), type = "both") plotInitialGuess(scores(sce, "doublet"), type = "truncated")
Read in a dataset and prepare it for analysis
readCytof( file.name, beads = c("Bead"), dna = c("DNA1", "DNA2"), event_length = "Event_length", viability = "Live_Dead", gaussian = c("Center", "Offset", "Width", "Residual"), verbose = TRUE )readCytof( file.name, beads = c("Bead"), dna = c("DNA1", "DNA2"), event_length = "Event_length", viability = "Live_Dead", gaussian = c("Center", "Offset", "Width", "Residual"), verbose = TRUE )
file.name |
A path to an .fcs file that contains CyTOF data or a
|
beads |
character vector that contains the names of all of the bead channels. |
dna |
Character vector that contains the names of the DNA markers. |
event_length |
Character vector of the event length variable. |
viability |
Character vector of the permeability/viability markers. |
gaussian |
Character vector that contains the names of the Gaussian Discrimination Parameters. |
verbose |
Logical value indicating whether or not to print a summary of the technical channels identified in the data. |
The function returns a SingleCellExperiment that contains all of
the original information from the fcs file. The data are imported using
CATALYST and then information is added to the colData
that will be used to determine labels for each event and to provide
additional information about the events that can be used for
exploratory data analysis and to aid the user in labeling the data.
The objects are all initialized at this point an values are filled
in during later stages of the labeling process.
Note that the names from the fcs file are required as arguments
to the readCytof. If you are not sure what those names are,
there is some code in the example that shows how to import your
data into a SingleCellExperiment using prepData from
CATALYST and look at the names.
A SingleCellExperiment that contains the information from
the CyTOF fcs file, the technical data that will be used to label
the data, and other objects that are used to store information through
the labeling process. The objects are DataFrame objects that
are stored in the colData for the SingleCellExperiment.
The objects are:
label |
A single variable |
probs |
A |
tech |
A |
scores |
Scores are computed to determine how much an event looks
like a bead, debris, doublet, or dead cell. These scores are used
to select a training dataset for the classification model, but they
can be helpful for exploratory data analysis so they are provided in
this |
initial |
Initial classification of each event type is determined
using a mixture model and the event type score. The |
library(CATALYST) library(SingleCellExperiment) data("raw_data", package = "CATALYST") # Determine at the names of the bead, DNA, and viability channels in the # file. Names are 'Beads', 'DNA1', 'DNA2', 'cisPt1', 'cisPt2'. tech <- prepData(raw_data) rownames(tech) # Determine names of event length and Gaussian parameters # names are 'Event_length', 'Center', 'Offset', 'Width', 'Residual' names(int_colData(tech)) # read in the data for use with cytofQC x <- readCytof(raw_data, beads = 'Beads', viability = c('cisPt1','cisPt2'))library(CATALYST) library(SingleCellExperiment) data("raw_data", package = "CATALYST") # Determine at the names of the bead, DNA, and viability channels in the # file. Names are 'Beads', 'DNA1', 'DNA2', 'cisPt1', 'cisPt2'. tech <- prepData(raw_data) rownames(tech) # Determine names of event length and Gaussian parameters # names are 'Event_length', 'Center', 'Offset', 'Width', 'Residual' names(int_colData(tech)) # read in the data for use with cytofQC x <- readCytof(raw_data, beads = 'Beads', viability = c('cisPt1','cisPt2'))
Returns the final label assignments for a parameter using a random forest
rfLabel( x, type = c("bead", "doublet", "debris", "dead"), loss = c("auc", "class"), n = 4000, standardize = TRUE )rfLabel( x, type = c("bead", "doublet", "debris", "dead"), loss = c("auc", "class"), n = 4000, standardize = TRUE )
x |
A |
type |
Identifies the type of label that is being modeled. Must be 'bead', 'doublet', 'debris', or 'dead'. |
loss |
Specifies the type of loss used to tune the GBM. Can be either "auc" for the area under the curve or "class" for classification error. |
n |
number of observations in training dataset. |
standardize |
Indicates if the data should be standardized. Because the data are on different scales, it should be standardized for this analysis because the variables are on different scales. |
rfLabel uses a random forest to compute the final labels
for the specified parameter type (bead, doublet, debris, or dead). This step
cannot be completed until the corresponding initialization function
(initialBead, initialDebris, initialDoublet, or
initialDead) is done on the SingleCellExperiment created by
readCytof.
The random forest uses the defaults from randomForest and then
predicted values are computed for all of the events in x. If the
predicted probability for the label type is greater than 0.5, the label is
changed to the specified type. However, if an observation already has a
label other than 'cell' in the label variable, it will not be changed.
The predicted probabilities for all of the observations are stored in the
variable associated with that type in the probs object of x
for further analysis. Thus, it is possible to have a probability greater
than 0.5 for 'debris' but still have a label of 'bead' if an observation
was classified as a bead prior to classifying the debris.
An updated SingleCellExperiment is returned with the labels
for the parameter of interest (bead, doublet, debris, or dead) added to
the label object of the SingleCellExperiment and the
probabilities for the event type added to the probs object of the
SingleCellExperiment.
data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) sce <- initialBead(sce) sce <- rfLabel(sce, type = "bead") head(probs(sce)) table(label(sce))data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) sce <- initialBead(sce) sce <- rfLabel(sce, type = "bead") head(probs(sce)) table(label(sce))
Returns the final label assignments for a parameter using a semi-supervised support vector machine
s3vmLabel( x, type = c("bead", "doublet", "debris", "dead"), loss = c("auc", "class"), n = 4000, standardize = TRUE )s3vmLabel( x, type = c("bead", "doublet", "debris", "dead"), loss = c("auc", "class"), n = 4000, standardize = TRUE )
x |
A |
type |
Identifies the type of label that is being modeled. Must be 'bead', 'doublet', 'debris', or 'dead'. |
loss |
Specifies the type of loss used to tune the SVM. Can be either "auc" for the area under the curve or "class" for classification error. |
n |
number of observations in training dataset. |
standardize |
Indicates if the data should be standardized. Because the data are on different scales, it should be standardized for this analysis because the variables are on different scales. |
s3vmLabel uses a semi-supervised support vector machine to
compute the final labels for the specified parameter type (bead, doublet,
debris, or dead). The model is initially computed using only the data
specified in the index argument. Events are iteratively added to this set
when the updated SVM predicts a label with high confidence. Then predicted
values are computed for all of the observations in x. If the
predicted probability for the label type is greater than 0.5, the label is
changed to the specified type. However, if an observation already has a
label other than 'cell' in the labels$label variable, it will not be
changed. The predicted probabilities for all of the observations is stored
in the variable associated with that type for further analysis. Thus, it is
possible to have a probability greater than 0.5 for 'debris' but still have
a label of 'bead' if an observation was classified as a bead prior to
classifying the debris.
An updated SingleCellExperiment is returned with the labels
for the parameter of interest (bead, doublet, debris, or dead) added to
the label object of the SingleCellExperiment and the
probabilities for the event type added to the probs object of the
SingleCellExperiment.
data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) sce <- initialBead(sce) sce <- svmLabel(sce, type = "bead", loss = "auc") head(probs(sce)) table(label(sce))data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) sce <- initialBead(sce) sce <- svmLabel(sce, type = "bead", loss = "auc") head(probs(sce)) table(label(sce))
Returns a specified object from the cytofQC SingleCellExperiment
scores(x, type = c("all", "bead", "debris", "doublet", "dead")) probs(x, type = c("all", "bead", "debris", "doublet", "dead")) label(x) tech( x, type = c("all", "Bead", "DNA", "Viability", "Event_length", "Center", "Offset", "Width", "Residual") ) initial(x, type = c("all", "bead", "debris", "doublet", "dead"))scores(x, type = c("all", "bead", "debris", "doublet", "dead")) probs(x, type = c("all", "bead", "debris", "doublet", "dead")) label(x) tech( x, type = c("all", "Bead", "DNA", "Viability", "Event_length", "Center", "Offset", "Width", "Residual") ) initial(x, type = c("all", "bead", "debris", "doublet", "dead"))
x |
A |
type |
Identifies the type of objects to be returned. For
|
For probs, scores, and tech, a numeric vector
or DataFrame with the information for the event type(s) is returned.
For label, a character vector containing the label for each event is
returned.
For tech, a DataFrame containing the technical variables
used to determine the label of each event. The bead, DNA, and viability
variables have an arcsinh transform, Event_length is unchanged, and
the Gaussian parameters have a log transform using log1p.
data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) sce <- labelQC(sce) table(label(sce)) cytofHist(scores(sce, type = 'bead'), label(sce), title = "Bead score")data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) sce <- labelQC(sce) table(label(sce)) cytofHist(scores(sce, type = 'bead'), label(sce), title = "Bead score")
Returns the final label assignments for a parameter using a support vector machine
svmLabel( x, type = c("bead", "doublet", "debris", "dead"), loss = c("auc", "class"), n = 4000, standardize = TRUE )svmLabel( x, type = c("bead", "doublet", "debris", "dead"), loss = c("auc", "class"), n = 4000, standardize = TRUE )
x |
A |
type |
Identifies the type of label that is being modeled. Must be 'bead', 'doublet', 'debris', or 'dead'. |
loss |
Specifies the type of loss used to tune the GBM. Can be either "auc" for the area under the curve or "class" for classification error. |
n |
number of observations in training dataset. |
standardize |
Indicates if the data should be standardized. Because the data are on different scales, it should be standardized for this analysis because the variables are on different scales. |
svmLabel uses a support vector machine to compute the final labels
for the specified parameter type (bead, doublet, debris, or dead). This step
cannot be completed until the corresponding initialization function
(initialBead, initialDebris, initialDoublet, or
initialDead) is done on the SingleCellExperiment created by
readCytof.
The support vector machine is tuned using eztune and
then predicted values are computed for all of the events in x. If
the predicted probability for the label type is greater than 0.5, the label
is changed to the specified type. However, if an observation already has a
label other than 'cell' in the label variable, it will not be
changed. The predicted probabilities for all of the observations are
stored in the variable associated with that type in the probs
object of x for further analysis. Thus, it is possible to have
a probability greater than 0.5 for 'debris' but still have a label of
'bead' if an observation was classified as a bead prior to classifying
the debris.
An updated SingleCellExperiment is returned with the labels
for the parameter of interest (bead, doublet, debris, or dead) added to
the label object of the SingleCellExperiment and the
probabilities for the event type added to the probs object of the
SingleCellExperiment.
data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) sce <- initialBead(sce) sce <- svmLabel(sce, type = "bead", loss = "auc") table(label(sce))data("raw_data", package = "CATALYST") sce <- readCytof(raw_data, beads = "Beads", viability = c("cisPt1", "cisPt2")) sce <- initialBead(sce) sce <- svmLabel(sce, type = "bead", loss = "auc") table(label(sce))