Skip to content

Commit

Permalink
minor amends to plotting intensity
Browse files Browse the repository at this point in the history
  • Loading branch information
@carolinesands
carolinesands committed Mar 20, 2024
1 parent bdf8192 commit ceecb60
Show file tree
Hide file tree
Showing 2 changed files with 88 additions and 20 deletions.
97 changes: 83 additions & 14 deletions nPYc/plotting/_multivariatePlotting.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -25,6 +25,7 @@
from ..utilities.generic import createDestinationPath
import copy
import datetime
from ..enumerations import AssayRole, SampleType


def plotScree(R2, Q2=None, title='', xlabel='', ylabel='', savePath=None, figureFormat='png', dpi=72,
Expand Down Expand Up @@ -560,8 +561,18 @@ def plotLoadings(pcaModel, msData, title='', figures=None, savePath=None, figure
return figures


def plotScoresInteractive(dataset, pcaModel, colourBy, components=[1, 2], alpha=0.05, withExclusions=False,
destinationPath=None, autoOpen=True):
def plotScoresInteractive(dataset,
pcaModel,
labelBy='Run Order',
colourBy='Run Order',
colourDict=None,
markerDict=None,
components=[1, 2],
alpha=0.05,
withExclusions=False,
destinationPath=None,
autoOpen=True,
opacity=.6):
"""
Interactively visualise PCA scores (coloured by a given sampleMetadata field, and for a given pair of components) with plotly, provides tooltips to allow identification of samples.
Expand All @@ -585,6 +596,9 @@ def plotScoresInteractive(dataset, pcaModel, colourBy, components=[1, 2], alpha=
values = pcaModel.scores
ns, nc = values.shape

if labelBy not in dataset.sampleMetadata.columns:
raise ValueError('labelBy must be a column in dataset.sampleMetadata')

if colourBy not in dataset.sampleMetadata.columns:
raise ValueError('colourBy must be a column in dataset.sampleMetadata')

Expand Down Expand Up @@ -619,10 +633,15 @@ def plotScoresInteractive(dataset, pcaModel, colourBy, components=[1, 2], alpha=

# Data preparation
classes = dataMasked.sampleMetadata[colourBy]
hovertext = dataMasked.sampleMetadata['Sample File Name'].str.cat(classes.astype(str), sep='; ' + colourBy + ': ')
hovertext = dataMasked.sampleMetadata['Sample File Name'].str.cat(dataMasked.sampleMetadata[labelBy].astype(str), sep='; ' + colourBy + ': ')
plotnans = classes.isnull().values
data = []

# If colourBy=='SampleClass' - NPC derived classes, then use default colours and markers
if colourBy == 'SampleClass':
colourDict = dataset.Attributes['sampleTypeColours']
markerDict = dataset.Attributes['sampleTypeMarkers']

# Ensure all values in column have the same type

# list of all types in column; and set of unique types
Expand Down Expand Up @@ -675,22 +694,72 @@ def plotScoresInteractive(dataset, pcaModel, colourBy, components=[1, 2], alpha=
# Plot categorical values by unique groups
else:
uniq = numpy.unique(classes[plotnans == False])

if colourDict is None:
colourDict={}
colors = iter(plt.cm.rainbow(numpy.linspace(0, 1, len(uniq))))
for u in uniq:
colourDict[u] = rgb2hex(next(colors))

if markerDict is None:
markerDict = {}
for u in uniq:
markerDict[u] = 'diamond-dot'

for i in uniq:
CLASSplot = go.Scatter(
x=values[classes == i, components[0]],
y=values[classes == i, components[1]],
mode='markers',
marker=dict(
colorscale='Portland',
symbol='circle',
),
text=hovertext[classes == i],
name=i,
hoverinfo='text',
showlegend=True
x=values[classes == i, components[0]],
y=values[classes == i, components[1]],
mode='markers',
marker=dict(
color=colourDict[i],
symbol=markerDict[i]
),
text=hovertext[classes == i],
name=str(i),
hoverinfo='text',
showlegend=True
)
data.append(CLASSplot)

# Overlay SR and LTR if columns present
if (colourBy != 'SampleClass') & ('SampleType' in dataMasked.sampleMetadata.columns) & ('AssayRole' in dataMasked.sampleMetadata.columns):
SRmask = ((dataMasked.sampleMetadata['SampleType'].values == SampleType.StudyPool) &
(dataMasked.sampleMetadata['AssayRole'].values == AssayRole.PrecisionReference))
LTRmask = ((dataMasked.sampleMetadata['SampleType'].values == SampleType.ExternalReference) &
(dataMasked.sampleMetadata['AssayRole'].values == AssayRole.PrecisionReference))

SRplot = go.Scatter(
x=values[SRmask, components[0]],
y=values[SRmask, components[1]],
mode='markers',
marker=dict(
color='darkgreen',
symbol='1',
),
text=hovertext[SRmask],
name='Study Reference',
hoverinfo='text',
showlegend=True,
)
data.append(SRplot)

LTRplot = go.Scatter(
x=values[LTRmask, components[0]],
y=values[LTRmask, components[1]],
mode='markers',
marker=dict(
color='darkorange',
symbol='2',

),
text=hovertext[LTRmask],
name='Long-Term Reference',
hoverinfo='text',
showlegend=True
)
data.append(LTRplot)

hotelling_ellipse = pcaModel.hotelling_T2(comps=numpy.array([components[0], components[1]]), alpha=alpha)

layout = {
Expand Down
11 changes: 5 additions & 6 deletions nPYc/plotting/_plotTIC.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -399,9 +399,7 @@ def plotIntensityInteractive(dataset,
for u in uniq:
colourDict[u] = rgb2hex(next(colors))

#colourDict = {}
#for u in uniq:
# colourDict[u] = 'magenta'
if markerDict is None:

markerDict = {}
for u in uniq:
Expand All @@ -419,7 +417,8 @@ def plotIntensityInteractive(dataset,
text=hovertext[classes == i],
name=str(i),
hoverinfo='text',
showlegend=True, opacity=opacity
showlegend=True,
opacity=opacity
)
data.append(CLASSplot)

Expand All @@ -441,7 +440,7 @@ def plotIntensityInteractive(dataset,
text=hovertext[SRmask],
name='Study Reference',
hoverinfo='text',
showlegend=True, opacity=opacity
showlegend=True
)
data.append(SRplot)

Expand All @@ -457,7 +456,7 @@ def plotIntensityInteractive(dataset,
text=hovertext[LTRmask],
name='Long-Term Reference',
hoverinfo='text',
showlegend=True, opacity=opacity
showlegend=True
)
data.append(LTRplot)

Expand Down

0 comments on commit ceecb60

Please sign in to comment.