def scatterplot_tab():
#File input
record_dict = SeqIO.index(
join(dirname(__file__), "data/protein_sequences.fasta"), "fasta")
df = pd.read_csv(join(dirname(__file__), 'data/input_data.csv'),
sep="\t",
header=0)
df = df.fillna(0.0)
cut_off = 2.0
output_backend = "webgl"
def create_figure(cut_off, df, output_backend):
columns = sorted(df.columns[2:])
#Initialize variables for data input
xs = df[x.value].values
ys = df[y.value].values
geneid = df['AlternateID'].values
desc = df['Annotation'].values
x_title = x.value.lower()
y_title = y.value.lower()
#create a new column in df named seq storing protein sequences
result = {}
for i in geneid:
if i in record_dict.keys():
result[i] = str(record_dict[i].seq)
else:
result[i] = str('*')
df['seq'] = result.values()
seq = df['seq'].values
data = {
'x_values':
xs,
'y_values':
ys,
'fold_change': [(-x / y) if (x > y and y != 0) else (y / x) if
(y >= x and x != 0) else 'NaN'
for x, y in zip(xs, ys)],
'GeneID':
geneid,
'Annotation':
desc,
'seq':
seq
}
mysource = ColumnDataSource(data=data)
#use CDSView function to determine a subset of significant genes for visualization
significant_xy_view = CDSView(
source=mysource,
filters=[
BooleanFilter([
True if
((geneX > 0 and geneY > 0 and
(abs(geneX / geneY) > cut_off)) or
(geneX > 0 and geneY > 0 and abs(geneY / geneX) > cut_off)
and (geneX + geneY) > 10) else False
for geneX, geneY in zip(mysource.data['x_values'],
mysource.data['y_values'])
])
])
# create a plot and style its properties
kw = dict()
kw['title'] = "%s vs %s" % (x_title, y_title)
#make a plot
p = figure(plot_height=500,
plot_width=700,
output_backend=output_backend,
tools='pan,box_zoom,wheel_zoom,reset,save,tap',
**kw)
#make sure WheelZoomTool is activated by default
p.toolbar.active_scroll = p.select_one(WheelZoomTool)
p.xaxis.axis_label = x_title
p.yaxis.axis_label = y_title
p.yaxis.axis_label_text_font_size = '10pt'
p.xaxis.axis_label_text_font_size = '10pt'
p.yaxis.major_label_text_font_size = '10pt'
p.xaxis.major_label_text_font_size = '10pt'
p.title.text_font_size = '12pt'
common_circle_kwargs = {
'x': 'x_values',
'y': 'y_values',
'source': mysource,
'hover_color': 'black'
}
p.circle(**common_circle_kwargs,
color="skyblue",
alpha=0.7,
size=9,
line_color="white")
p.circle(**common_circle_kwargs,
view=significant_xy_view,
color="red",
muted_alpha=0.1,
legend_label='sig. genes',
size=9,
line_color="black")
# Highlighting of genes using a substring in the gene annotation, if it is provided in the Textfield
if desc_sel.value != "":
selected = df[df.Annotation.str.contains(desc_sel.value,
case=False) == True]
p.circle(x=selected[x.value].values,
y=selected[y.value].values,
color="yellow",
alpha=1.0,
muted_alpha=0.1,
legend_label='sel. genes',
size=8,
line_color="black")
# Set autohide to true to only show the toolbar when mouse is over plot
p.toolbar.autohide = True
p.legend.title = 'Click to hide'
p.legend.location = "bottom_right"
p.legend.border_line_width = 1
p.legend.border_line_color = "black"
p.legend.border_line_alpha = 0.5
# use the "seq" column of mysource to complete the URL
# e.g. if the glyph at index 1 is selected, then @seq
# will be replaced with mysource.data['seq'][1]
url = "http://papers.genomics.lbl.gov/cgi-bin/litSearch.cgi?query=@seq&Search=Search"
taptool = p.select(type=TapTool)
taptool.callback = OpenURL(url=url)
# Format the tooltip
tooltips = [('GeneID', '@GeneID'), ('Annotation', '@Annotation{safe}'),
('fold change', '@fold_change'),
('FPKM (%s)' % (x_title), '@x_values'),
('FPKM (%s)' % (y_title), '@y_values')]
# Configure a renderer to be used upon hover
hover_glyph = p.circle(**common_circle_kwargs,
size=15,
alpha=0,
hover_fill_color='black',
hover_alpha=0.5)
# Add the custom HoverTool to the figure
p.add_tools(HoverTool(tooltips=tooltips, renderers=[hover_glyph]))
# Add the custom CrosshairTool to the figure
p.add_tools(CrosshairTool(line_width=1))
# Add interactivity to the legend
p.legend.click_policy = "hide"
return p
def update_data(attr, old, new):
cut_off = float(sl.value)
layout.children[1] = create_figure(cut_off, df, output_backend)
x = Select(title='X-Axis: condition A',
value='wt_0min',
options=sorted(list(df.iloc[:, 2:].columns)))
x.on_change('value', update_data)
y = Select(title='Y-Axis: condition B',
value='wt_25min',
options=sorted(list(df.iloc[:, 2:].columns)))
y.on_change('value', update_data)
sl = Slider(start=0.0,
end=10.0,
value=cut_off,
step=0.5,
title='|fold change| \u003E')
sl.callback_policy = 'mouseup'
sl.on_change('value_throttled', update_data)
desc_sel = TextInput(title="Gene annotation contains:")
desc_sel.on_change('value', update_data)
div = Div(text="""
Interact with the widgets below to query a list of significant genes to plot.
Hover over the circles to see more information about each gene.
Click on the circle to get information from PaperBlast""")
#add button for saving plot as .svg
def button_handler():
timestr = time.strftime("%Y%m%d-%H%M%S")
p.output_backend = 'svg'
export_svgs(p, filename="scatterplot{}.svg".format(timestr))
p.output_backend = 'webgl'
button = Button(label="Save as .svg", button_type="success")
button.on_click(button_handler)
p = create_figure(cut_off, df, output_backend)
# Make a tab with the layout
controls = column(x, y, sl, desc_sel, button, width=200)
layout = row(column(div, controls), p)
tab = pn.pane.Bokeh(layout)
return tab
def slider_cb(attr, old, new):
# call draw function and put the new figure in the layout
p = illusion.draw(permMap[variation_selector.active],
distortion_slider.value)
pBox.children[0] = p
submit_button.on_click(submit_button_cb)
variation_selector.on_change('active', selector_cb)
save_button.on_click(save_button_cb)
#slider.on_change('value', cb)
# hack to throttle slider callback (https://stackoverflow.com/questions/38375961/throttling-in-bokeh-application/38379136#38379136)
distortion_slider.callback_policy = 'mouseup' #call only on mouseup
#slider.callback_throttle = 50 #call max every x ms
source = ColumnDataSource(data=dict(value=[]))
source.on_change('data', slider_cb)
distortion_slider.callback = CustomJS(args=dict(source=source),
code="""
source.data = { value: [cb_obj.value] }
""")
# ## some CSS to center layout
# header = Div(text="""
# <style>
# body { width: 1150px; margin: 0 auto; }
# </style>
# """)
# curdoc().add_root(header)
