def plot(
self,
figsize: Tuple[float, float] = (15, 10),
same_plot: bool = False,
hide_cells: bool = False,
perc: Tuple[float, float] = None,
abs_prob_cmap: mcolors.ListedColormap = cm.viridis,
cell_color: str = "black",
lineage_color: str = "black",
alpha: float = 0.8,
lineage_alpha: float = 0.2,
title: Optional[str] = None,
size: int = 15,
lw: float = 2,
show_cbar: bool = True,
margins: float = 0.015,
xlabel: str = "pseudotime",
ylabel: str = "expression",
show_conf_int: bool = True,
dpi: int = None,
fig: mpl.figure.Figure = None,
ax: mpl.axes.Axes = None,
return_fig: bool = False,
save: Optional[str] = None,
) -> Optional[mpl.figure.Figure]:
"""
Plot the smoothed gene expression.
Parameters
----------
figsize
Size of the figure.
same_plot
Whether to plot all trends in the same plot.
hide_cells
Whether to hide the cells.
perc
Percentile by which to clip the absorption probabilities./
abs_prob_cmap
Colormap to use when coloring in the absorption probabilities.
cell_color
Color for the cells when not coloring absorption probabilities.
lineage_color
Color for the lineage.
alpha
Alpha channel for cells.
lineage_alpha
Alpha channel for lineage confidence intervals.
title
Title of the plot.
size
Size of the points.
lw
Line width for the smoothed values.
show_cbar
Whether to show colorbar.
margins
Margins around the plot.
xlabel
Label on the x-axis.
ylabel
Label on the y-axis.
show_conf_int
Whether to show the confidence interval.
dpi
Dots per inch.
fig
Figure to use, if `None`, create a new one.
ax: :class:`matplotlib.axes.Axes`
Ax to use, if `None`, create a new one.
return_fig
If `True`, return the figure object.
save
Filename where to save the plot. If `None`, just shows the plots.
Returns
-------
%(just_plots)s
"""
if fig is None or ax is None:
fig, ax = plt.subplots(figsize=figsize, constrained_layout=True)
if dpi is not None:
fig.set_dpi(dpi)
vmin, vmax = _minmax(self.w, perc)
if not hide_cells:
_ = ax.scatter(
self.x_all.squeeze(),
self.y_all.squeeze(),
c=cell_color if same_plot or np.allclose(self.w_all, 1.0) else
self.w_all.squeeze(),
s=size,
cmap=abs_prob_cmap,
vmin=vmin,
vmax=vmax,
alpha=alpha,
)
if title is None:
title = f"{self._gene} @ {self._lineage}"
_ = ax.plot(self.x_test,
self.y_test,
color=lineage_color,
lw=lw,
label=title)
ax.set_title(title)
ax.set_ylabel(ylabel)
ax.set_xlabel(xlabel)
ax.margins(margins)
if show_conf_int and self.conf_int is not None:
ax.fill_between(
self.x_test.squeeze(),
self.conf_int[:, 0],
self.conf_int[:, 1],
alpha=lineage_alpha,
color=lineage_color,
linestyle="--",
)
if (show_cbar and not hide_cells and not same_plot
and not np.allclose(self.w_all, 1)):
norm = mcolors.Normalize(vmin=vmin, vmax=vmax)
divider = make_axes_locatable(ax)
cax = divider.append_axes("right", size="2.5%", pad=0.1)
_ = mpl.colorbar.ColorbarBase(cax,
norm=norm,
cmap=abs_prob_cmap,
label="absorption probability")
if save is not None:
save_fig(fig, save)
if return_fig:
return fig
def plot(
self,
figsize: Tuple[float, float] = (8, 5),
same_plot: bool = False,
hide_cells: bool = False,
perc: Tuple[float, float] = None,
abs_prob_cmap: mcolors.ListedColormap = cm.viridis,
cell_color: str = "black",
lineage_color: str = "black",
alpha: float = 0.8,
lineage_alpha: float = 0.2,
title: Optional[str] = None,
size: int = 15,
lw: float = 2,
cbar: bool = True,
margins: float = 0.015,
xlabel: str = "pseudotime",
ylabel: str = "expression",
conf_int: bool = True,
lineage_probability: bool = False,
lineage_probability_conf_int: Union[bool, float] = False,
lineage_probability_color: Optional[str] = None,
dpi: int = None,
fig: mpl.figure.Figure = None,
ax: mpl.axes.Axes = None,
return_fig: bool = False,
save: Optional[str] = None,
**kwargs,
) -> Optional[mpl.figure.Figure]:
"""
Plot the smoothed gene expression.
Parameters
----------
figsize
Size of the figure.
same_plot
Whether to plot all trends in the same plot.
hide_cells
Whether to hide the cells.
perc
Percentile by which to clip the absorption probabilities.
abs_prob_cmap
Colormap to use when coloring in the absorption probabilities.
cell_color
Color for the cells when not coloring absorption probabilities.
lineage_color
Color for the lineage.
alpha
Alpha channel for cells.
lineage_alpha
Alpha channel for lineage confidence intervals.
title
Title of the plot.
size
Size of the points.
lw
Line width for the smoothed values.
cbar
Whether to show colorbar.
margins
Margins around the plot.
xlabel
Label on the x-axis.
ylabel
Label on the y-axis.
conf_int
Whether to show the confidence interval.
lineage_probability
Whether to show smoothed lineage probability as a dashed line.
Note that this will require 1 additional model fit.
lineage_probability_conf_int
Whether to compute and show smoothed lineage probability confidence interval.
If :paramref:`self` is :class:`cellrank.ul.models.GAMR`, it can also specify the confidence level,
the default is `0.95`. Only used when ``show_lineage_probability=True``.
lineage_probability_color
Color to use when plotting the smoothed ``lineage_probability``.
If `None`, it's the same as ``lineage_color``. Only used when ``show_lineage_probability=True``.
dpi
Dots per inch.
fig
Figure to use, if `None`, create a new one.
ax: :class:`matplotlib.axes.Axes`
Ax to use, if `None`, create a new one.
return_fig
If `True`, return the figure object.
save
Filename where to save the plot. If `None`, just shows the plots.
**kwargs
Keyword arguments for :meth:`matplotlib.axes.Axes.legend`, e.g. to disable the legend, specify ``loc=None``.
Only available when ``show_lineage_probability=True``.
Returns
-------
%(just_plots)s
"""
if self.y_test is None:
raise RuntimeError("Run `.predict()` first.")
if fig is None or ax is None:
fig, ax = plt.subplots(figsize=figsize, constrained_layout=True)
if dpi is not None:
fig.set_dpi(dpi)
conf_int = conf_int and self.conf_int is not None
hide_cells = (hide_cells or self.x_all is None or self.w_all is None
or self.y_all is None)
lineage_probability_color = (lineage_color
if lineage_probability_color is None else
lineage_probability_color)
scaler = kwargs.pop(
"scaler",
self._create_scaler(
lineage_probability,
show_conf_int=conf_int,
),
)
if lineage_probability:
if ylabel in ("expression", self._gene):
ylabel = f"scaled {ylabel}"
vmin, vmax = None, None
if not hide_cells:
vmin, vmax = _minmax(self.w_all, perc)
_ = ax.scatter(
self.x_all.squeeze(),
scaler(self.y_all.squeeze()),
c=cell_color if same_plot or np.allclose(self.w_all, 1.0) else
self.w_all.squeeze(),
s=size,
cmap=abs_prob_cmap,
vmin=vmin,
vmax=vmax,
alpha=alpha,
)
if title is None:
title = (f"{self._gene} @ {self._lineage}"
if self._lineage is not None else f"{self._gene}")
ax.plot(self.x_test,
scaler(self.y_test),
color=lineage_color,
lw=lw,
label=title)
if title is not None:
ax.set_title(title)
if ylabel is not None:
ax.set_ylabel(ylabel)
if xlabel is not None:
ax.set_xlabel(xlabel)
ax.margins(margins)
if conf_int:
ax.fill_between(
self.x_test.squeeze(),
scaler(self.conf_int[:, 0]),
scaler(self.conf_int[:, 1]),
alpha=lineage_alpha,
color=lineage_color,
linestyle="--",
)
if (lineage_probability and not isinstance(self, FittedModel)
and not np.allclose(self.w, 1.0)):
from cellrank.pl._utils import _is_any_gam_mgcv
model = deepcopy(self)
model._y = self._reshape_and_retype(self.w).copy()
model = model.fit()
if not lineage_probability_conf_int:
y = model.predict()
elif _is_any_gam_mgcv(model):
y = model.predict(
level=lineage_probability_conf_int if isinstance(
lineage_probability_conf_int, float) else 0.95)
else:
y = model.predict()
model.confidence_interval()
ax.fill_between(
model.x_test.squeeze(),
model.conf_int[:, 0],
model.conf_int[:, 1],
alpha=lineage_alpha,
color=lineage_probability_color,
linestyle="--",
)
handle = ax.plot(
model.x_test,
y,
color=lineage_probability_color,
lw=lw,
linestyle="--",
zorder=-1,
label="probability",
)
if kwargs.get("loc", "best") is not None:
ax.legend(handles=handle, **kwargs)
if (cbar and not hide_cells and not same_plot
and not np.allclose(self.w_all, 1.0)):
norm = mcolors.Normalize(vmin=vmin, vmax=vmax)
divider = make_axes_locatable(ax)
cax = divider.append_axes("right", size="2%", pad=0.1)
_ = mpl.colorbar.ColorbarBase(
cax,
norm=norm,
cmap=abs_prob_cmap,
ticks=np.linspace(norm.vmin, norm.vmax, 5),
)
if save is not None:
save_fig(fig, save)
if return_fig:
return fig