cytoflow.operations.gaussian¶
-
class
cytoflow.operations.gaussian.
GaussianMixtureOp
[source]¶ Bases:
traits.has_traits.HasStrictTraits
This module fits a Gaussian mixture model with a specified number of components to one or more channels.
If
num_components
> 1
,apply()
creates a new categorical metadata variable namedname
, with possible values{name}_1
….name_n
wheren
is the number of components. An event is assigned toname_i
category if it has the highest posterior probability of having been produced by componenti
. If an event has a value that is outside the range of one of the channels’ scales, then it is assigned to{name}_None
.Optionally, if
sigma
is greater than 0,apply()
creates newboolean
metadata variables named{name}_1
…{name}_n
wheren
is the number of components. The column{name}_i
isTrue
if the event is less thansigma
standard deviations from the mean of componenti
. Ifnum_components
is1
,sigma
must be greater than 0.Optionally, if
posteriors
isTrue
,apply()
creates a newdouble
metadata variables named{name}_1_posterior
…{name}_n_posterior
wheren
is the number of components. The column{name}_i_posterior
contains the posterior probability that this event is a member of componenti
.Finally, the same mixture model (mean and standard deviation) may not be appropriate for every subset of the data. If this is the case, you can use the
by
attribute to specify metadata by which to aggregate the data before estimating (and applying) a mixture model. The number of components must be the same across each subset, though.-
name
¶ The operation name; determines the name of the new metadata column
Type: Str
-
channels
¶ The channels to apply the mixture model to.
Type: List(Str)
-
scale
¶ Re-scale the data in the specified channels before fitting. If a channel is in
channels
but not inscale
, the current package-wide default (set withset_default_scale()
) is used.Type: Dict(Str : {“linear”, “logicle”, “log”})
-
num_components
¶ How many components to fit to the data? Must be a positive integer.
Type: Int (default = 1)
-
sigma
¶ How many standard deviations on either side of the mean to include in the boolean variable
{name}_i
? Must be>= 0.0
. Ifnum_components
is1
, must be> 0
.Type: Float (default = 0.0)
-
by
¶ A list of metadata attributes to aggregate the data before estimating the model. For example, if the experiment has two pieces of metadata,
Time
andDox
, settingby
to["Time", "Dox"]
will fit the model separately to each subset of the data with a unique combination ofTime
andDox
.Type: List(Str)
-
posteriors
¶ If
True
, add columns named{name}_{i}_posterior
giving the posterior probability that the event is in componenti
. Useful for filtering out low-probability events.Type: Bool (default = False)
Notes
We use the Mahalnobis distance as a multivariate generalization of the number of standard deviations an event is from the mean of the multivariate gaussian. If \(\vec{x}\) is an observation from a distribution with mean \(\vec{\mu}\) and \(S\) is the covariance matrix, then the Mahalanobis distance is \(\sqrt{(x - \mu)^T \cdot S^{-1} \cdot (x - \mu)}\).
Examples
Make a little data set.
>>> import cytoflow as flow >>> import_op = flow.ImportOp() >>> import_op.tubes = [flow.Tube(file = "Plate01/RFP_Well_A3.fcs", ... conditions = {'Dox' : 10.0}), ... flow.Tube(file = "Plate01/CFP_Well_A4.fcs", ... conditions = {'Dox' : 1.0})] >>> import_op.conditions = {'Dox' : 'float'} >>> ex = import_op.apply()
Create and parameterize the operation.
>>> gm_op = flow.GaussianMixtureOp(name = 'Gauss', ... channels = ['Y2-A'], ... scale = {'Y2-A' : 'log'}, ... num_components = 2)
Estimate the clusters
>>> gm_op.estimate(ex)
Plot a diagnostic view
>>> gm_op.default_view().plot(ex)
Apply the gate
>>> ex2 = gm_op.apply(ex)
Plot a diagnostic view with the event assignments
>>> gm_op.default_view().plot(ex2)
And with two channels:
>>> gm_op = flow.GaussianMixtureOp(name = 'Gauss', ... channels = ['V2-A', 'Y2-A'], ... scale = {'V2-A' : 'log', ... 'Y2-A' : 'log'}, ... num_components = 2) >>> gm_op.estimate(ex) >>> ex2 = gm_op.apply(ex) >>> gm_op.default_view().plot(ex2)
-
estimate
(experiment, subset=None)[source]¶ Estimate the Gaussian mixture model parameters
Parameters: - experiment (Experiment) – The data to use to estimate the mixture parameters
- subset (str (default = None)) – If set, a Python expression to determine the subset of the data to use to in the estimation.
-
apply
(experiment)[source]¶ Assigns new metadata to events using the mixture model estimated in
estimate()
.Returns: A new Experiment
with the new condition variables as described in the class documentation. Also adds the following new statistics:- mean : Float
- the mean of the fitted gaussian in each channel for each component.
- sigma : (Float, Float)
- the locations the mean +/- one standard deviation in each channel for each component.
- correlation : Float
- the correlation coefficient between each pair of channels for each component.
- proportion : Float
- the proportion of events in each component of the mixture model. only
added if
num_components
> 1
.
Return type: Experiment
-
-
class
cytoflow.operations.gaussian.
GaussianMixture1DView
[source]¶ Bases:
cytoflow.operations.base_op_views.By1DView
,cytoflow.operations.base_op_views.AnnotatingView
,cytoflow.views.histogram.HistogramView
A default view for
GaussianMixtureOp
that plots the histogram of a single channel, then the estimated Gaussian distributions on top of it.-
facets
¶ A read-only list of the conditions used to facet this view.
Type: List(String)
-
by
¶ A read-only list of the conditions used to group this view’s data before plotting.
Type: List(String)
-
channel
¶ The channel this view is viewing. If you created the view using
default_view()
, this is already set.Type: String
-
scale
¶ The way to scale the x axes. If you created the view using
default_view()
, this may be already set.Type: {‘linear’, ‘log’, ‘logicle’}
-
op
¶ The
IOperation
that this view is associated with. If you created the view usingdefault_view()
, this is already set.Type: Instance(IOperation)
-
xfacet, yfacet
Set to one of the
conditions
in theExperiment
, and a new row or column of subplots will be added for every unique value of that condition.Type: String
-
huefacet
¶ Set to one of the
conditions
in the in theExperiment
, and a new color will be added to the plot for every unique value of that condition.Type: String
-
plot
(experiment, **kwargs)[source]¶ Plot the plots.
Parameters: - experiment (Experiment) – The
Experiment
to plot using this view. - title (str) – Set the plot title
- xlabel, ylabel (str) – Set the X and Y axis labels
- huelabel (str) – Set the label for the hue facet (in the legend)
- legend (bool) – Plot a legend for the color or hue facet? Defaults to True.
- sharex, sharey (bool) – If there are multiple subplots, should they share axes? Defaults to True.
- row_order, col_order, hue_order (list) – Override the row/column/hue facet value order with the given list. If a value is not given in the ordering, it is not plotted. Defaults to a “natural ordering” of all the values.
- height (float) – The height of each row in inches. Default = 3.0
- aspect (float) – The aspect ratio of each subplot. Default = 1.5
- col_wrap (int) – If xfacet is set and yfacet is not set, you can “wrap” the subplots around so that they form a multi-row grid by setting col_wrap to the number of columns you want.
- sns_style ({“darkgrid”, “whitegrid”, “dark”, “white”, “ticks”}) – Which seaborn style to apply to the plot? Default is whitegrid.
- sns_context ({“paper”, “notebook”, “talk”, “poster”}) – Which seaborn context to use? Controls the scaling of plot elements such as tick labels and the legend. Default is talk.
- despine (Bool) – Remove the top and right axes from the plot? Default is True.
- min_quantile (float (>0.0 and <1.0, default = 0.001)) – Clip data that is less than this quantile.
- max_quantile (float (>0.0 and <1.0, default = 1.00)) – Clip data that is greater than this quantile.
- lim ((float, float)) – Set the range of the plot’s data axis.
- orientation ({‘vertical’, ‘horizontal’})
- num_bins (int) – The number of bins to plot in the histogram. Clipped to [100, 1000]
- histtype ({‘stepfilled’, ‘step’, ‘bar’}) – The type of histogram to draw. stepfilled is the default, which is a line plot with a color filled under the curve.
- density (bool) – If True, re-scale the histogram to form a probability density function, so the area under the histogram is 1.
- linewidth (float) – The width of the histogram line (in points)
- linestyle ([‘-’ | ‘–’ | ‘-.’ | ‘:’ | “None”]) – The style of the line to plot
- alpha (float (default = 0.5)) – The alpha blending value, between 0 (transparent) and 1 (opaque).
- color (matplotlib color) – The color to plot the annotations. Overrides the default color cycle.
- plot_name (str) – If this
IView
can make multiple plots,plot_name
is the name of the plot to make. Must be one of the values retrieved fromenum_plots()
.
- experiment (Experiment) – The
-
-
class
cytoflow.operations.gaussian.
GaussianMixture2DView
[source]¶ Bases:
cytoflow.operations.base_op_views.By2DView
,cytoflow.operations.base_op_views.AnnotatingView
,cytoflow.views.scatterplot.ScatterplotView
A default view for
GaussianMixtureOp
that plots the scatter plot of a two channels, then the estimated 2D Gaussian distributions on top of it.-
facets
¶ A read-only list of the conditions used to facet this view.
Type: List(String)
-
by
¶ A read-only list of the conditions used to group this view’s data before plotting.
Type: List(String)
-
xchannel, ychannel
The channels to use for this view’s X and Y axes. If you created the view using
default_view()
, this is already set.Type: String
-
xscale, yscale
The way to scale the x axes. If you created the view using
default_view()
, this may be already set.Type: {‘linear’, ‘log’, ‘logicle’}
-
op
¶ The
IOperation
that this view is associated with. If you created the view usingdefault_view()
, this is already set.Type: Instance(IOperation)
-
xlim, ylim
Set the min and max limits of the plots’ x and y axes.
Type: (float, float)
-
xfacet, yfacet
Set to one of the
conditions
in theExperiment
, and a new row or column of subplots will be added for every unique value of that condition.Type: String
-
huefacet
¶ Set to one of the
conditions
in the in theExperiment
, and a new color will be added to the plot for every unique value of that condition.Type: String
-
plot
(experiment, **kwargs)[source]¶ Plot the plots.
Parameters: - experiment (Experiment) – The
Experiment
to plot using this view. - title (str) – Set the plot title
- xlabel, ylabel (str) – Set the X and Y axis labels
- huelabel (str) – Set the label for the hue facet (in the legend)
- legend (bool) – Plot a legend for the color or hue facet? Defaults to True.
- sharex, sharey (bool) – If there are multiple subplots, should they share axes? Defaults to True.
- row_order, col_order, hue_order (list) – Override the row/column/hue facet value order with the given list. If a value is not given in the ordering, it is not plotted. Defaults to a “natural ordering” of all the values.
- height (float) – The height of each row in inches. Default = 3.0
- aspect (float) – The aspect ratio of each subplot. Default = 1.5
- col_wrap (int) – If xfacet is set and yfacet is not set, you can “wrap” the subplots around so that they form a multi-row grid by setting col_wrap to the number of columns you want.
- sns_style ({“darkgrid”, “whitegrid”, “dark”, “white”, “ticks”}) – Which seaborn style to apply to the plot? Default is whitegrid.
- sns_context ({“paper”, “notebook”, “talk”, “poster”}) – Which seaborn context to use? Controls the scaling of plot elements such as tick labels and the legend. Default is talk.
- despine (Bool) – Remove the top and right axes from the plot? Default is True.
- min_quantile (float (>0.0 and <1.0, default = 0.001)) – Clip data that is less than this quantile.
- max_quantile (float (>0.0 and <1.0, default = 1.00)) – Clip data that is greater than this quantile.
- xlim, ylim ((float, float)) – Set the range of the plot’s axis.
- alpha (float (default = 0.25)) – The alpha blending value, between 0 (transparent) and 1 (opaque).
- s (int (default = 2)) – The size in points^2.
- marker (a matplotlib marker style, usually a string) – Specfies the glyph to draw for each point on the scatterplot. See matplotlib.markers for examples. Default: ‘o’
- color (matplotlib color) – The color to plot the annotations. Overrides the default color cycle.
- plot_name (str) – If this
IView
can make multiple plots,plot_name
is the name of the plot to make. Must be one of the values retrieved fromenum_plots()
.
- experiment (Experiment) – The
-