Module omniplot.plot
Expand source code
from typing import Union, Optional, Dict, List
import matplotlib.collections as mc
import matplotlib.pyplot as plt
import numpy as np
import seaborn as sns
import pandas as pd
from matplotlib import cm
from matplotlib.lines import Line2D
from scipy.cluster.hierarchy import leaves_list
from scipy.cluster import hierarchy
from collections import defaultdict
import matplotlib.colors
from natsort import natsort_keygen, natsorted
from matplotlib.patches import Rectangle
import scipy.cluster.hierarchy as sch
import fastcluster as fcl
import sys
import matplotlib as mpl
from sklearn.cluster import KMeans, DBSCAN
from sklearn.metrics import silhouette_score
from scipy.spatial.distance import pdist, squareform
from sklearn.decomposition import PCA, NMF, LatentDirichletAllocation
from scipy.stats import fisher_exact
from scipy.stats import zscore
from itertools import combinations
import os
#script_dir = os.path.dirname( __file__ )
#sys.path.append( script_dir )
from omniplot.utils import * #
import scipy.stats as stats
from joblib import Parallel, delayed
from omniplot.chipseq_utils import _calc_pearson
import itertools as it
from omniplot.scatter import *
from omniplot.proportion import *
colormap_list: list=["nipy_spectral",
"terrain",
"tab20b",
"tab20c",
"gist_rainbow",
"hsv",
"CMRmap",
"coolwarm",
"gnuplot",
"gist_stern",
"brg",
"rainbow",
"jet"]
hatch_list: list = ['//', '\\\\', '||', '--', '++', 'xx', 'oo', 'OO', '..', '**','/o', '\\|', '|*', '-\\', '+o', 'x*', 'o-', 'O|', 'O.', '*-']
maker_list: list=['.', '_' , '+','|', 'x', 'v', '^', '<', '>', 's', 'p', '*', 'h', 'D', 'd', 'P', 'X','o', '1', '2', '3', '4','|', '_']
plt.rcParams['font.family']= 'sans-serif'
plt.rcParams['font.sans-serif'] = ['Arial']
plt.rcParams['svg.fonttype'] = 'none'
sns.set_theme(font="Arial")
def radialtree(df: pd.DataFrame,
n_clusters: int=3,
x: str="",
variables: List=[],
category: Union[str, List[str]]=[],
ztransform: bool=True,
save: str="",
distance_method="euclidean",
tree_method="ward",
title: str="",
y: list=[],linewidth: float=1,figsize: Optional[list]=None,
**kwargs) -> Dict:
"""
Drawing a radial dendrogram with color labels.
Parameters
----------
df : pandas DataFrame
A wide format data.
n_clusters: int, optional (default: 3)
Approximate number of clusters to produce
x: str, optional
the name of columns containing sample names. If not provided, the index will be considered sample names.
variables: list, optional
the name of columns containing variables to calculate the distances between samples
category: str or list of str
the column name of a category that is going to presented as colors around the dendrogram.
ztransform: bool=True,
save: str="",
distance_method="euclidean",
tree_method="ward",
show : bool
Whether or not to show the figure.
fontsize : float
A float to specify the font size
figsize : [x, y] array-like
1D array-like of floats to specify the figure size
palette : string
Matlab colormap name.
Returns
-------
dict: {"axes":ax, "clusters": clusters}
Raises
------
Notes
-----
References
----------
See Also
--------
Examples
--------
"""
if len(y)!=0:
variables=y
if x !="":
_labels=df[x]
df=df.drop(x, axis=1)
else:
_labels=df.index
X, category=_separate_data(df, variables=variables, category=category)
category_df=df[category]
# if len(variables)!=0 and len(category)!=0:
# if type(category)==str:
# category=[category]
# category_df=df[category]
# df=df[variables]
# X = df.values
# #print(X)
# assert X.dtype==float, f"{x} columns must contain only float values."
#
#
# elif len(category) !=0:
# if type(category)==str:
# category=[category]
# category_df=df[category]
# df=df.drop(category, axis=1)
# X = df.values
# #print(X)
# assert X.dtype==float, f"data must contain only float values except {category} column."
#
# else:
# X = df.values
# assert X.dtype==float, "data must contain only float values."
if ztransform==True:
X=zscore(X, axis=0)
D=squareform(pdist(X,metric=distance_method))
Y = sch.linkage(D, method=tree_method)
Z = sch.dendrogram(Y,labels=_labels,no_plot=True)
t=_dendrogram_threshold(Z, n_clusters)
_fig, _ax =plt.subplots()
Z=sch.dendrogram(Y,
labels = _labels,
color_threshold=t, ax=_ax)
#xticks=set(_ax.get_xticks())
plt.close(_fig)
sample_classes={k: list(category_df[k]) for k in category_df.columns}
ax=_radialtree2(Z,
sample_classes=sample_classes,
addlabels=False,
linewidth=linewidth,
figsize=figsize, **kwargs)
if title !="":
ax.set_title(title)
_save(save, "radialtree")
clusters = _get_cluster_classes(Z)
return {"axes":ax, "clusters":clusters}
def correlation(df: pd.DataFrame,
category: Union[str, list]=[],
variables: List=[],
method="pearson",
palette: str="coolwarm",
figsize=[6,6],
show_values=False,
clustermap_param:dict={},
ztransform: bool=True,
xticklabels =False,
yticklabels=False,
title: str="",)->Dict:
"""
Drawing a heatmap with correlations or distances between observations
Parameters
----------
df : pandas DataFrame
variables: List, optional
the names of values to calculate correlations
category: str or list, optional
the names of categorical values to display as color labels
mthod: str
method for correlation/distance calculation. Defalt: "pearson"
palette : str
A colormap name
show_values: bool, optional
Wheter to exhibit the values of fractions/counts/percentages.
clustermap_param : dict, optional
Whether or not to show the figure.
figsize : List[int], optional
The figure size, e.g., [4, 6].
ztransform : bool, optional
Whether to transform values to z-score
xticklabels, yticklabels : bool
Whether to show the label names in the heatmap
Returns
-------
dict
Raises
------
Notes
-----
References
----------
See Also
--------
Examples
--------
"""
original_index=df.index
X, category=_separate_data(df, variables=variables, category=category)
# if len(category) !=0:
#
# if type(category)==str:
# category=[category]
# #df=df.drop(category, axis=1)
# valnames=list(set(df.columns) -set(category))
# X = df[valnames].values
# assert X.dtype==float, f"data must contain only float values except {category} column."
#
# else:
# X = df.values
# assert X.dtype==float, "data must contain only float values."
if ztransform==True:
X=zscore(X, axis=0)
if method=="pearson":
# dmat=Parallel(n_jobs=-1)(delayed(_calc_pearson)(ind, X) for ind in list(it.combinations(range(X.shape[0]), 2)))
# dmat=np.array(dmat)
# dmat=squareform(dmat)
# print(dmat)
# dmat+=np.identity(dmat.shape[0])
dmat=np.corrcoef(X)
else:
dmat=squareform(pdist(X, method))
if method=="pearson":
ctitle="Pearson correlation"
else:
ctitle=method+" distance"
if len(category) >0:
dfm=pd.DataFrame(data=dmat)
colnames=dfm.columns
for cat in category:
dfm[cat]=df[cat].values
res=complex_clustermap(dfm,
heatmap_col=colnames,
row_colors=category,
ztranform=False,
xticklabels=xticklabels,
yticklabels=yticklabels,
figsize=figsize,
ctitle=ctitle )
else:
res=complex_clustermap(data=dmat,
xticklabels=xticklabels,
yticklabels=yticklabels,
method="ward",
cmap=palette,
col_cluster=True,
row_cluster=True,
figsize=figsize,
#rasterized=True,
#cbar_kws={"label":"Pearson correlation"},
#annot=show_values,
**clustermap_param)
return res
# g.cax.set_ylabel(ctitle, rotation=-90,va="bottom")
# plt.setp(g.ax_heatmap.get_yticklabels(), rotation=0) # For y axis
# plt.setp(g.ax_heatmap.get_xticklabels(), rotation=90) # For x axis
# return {"grid":g}
def triangle_heatmap(df,
grid_pos: list=[],
grid_labels: list=[],
show: bool=False,
save: str="",title: str="")-> dict:
"""
Creating a heatmap with 45 degree rotation.
Parameters
----------
df : pandas DataFrame
grid_pos: list
the column name of a category that is going to be placed in the row of the dotplot
grid_labels: list
the column name of a category that is going to be placed in the column of the dotplot
show : bool
Whether or not to show the figure.
Returns
-------
dict {"axes": ax}
Raises
------
Notes
-----
References
----------
See Also
--------
Examples
--------
"""
genes=df.index
fig, ax = plt.subplots(figsize=[8,6])
dmat=df.to_numpy()
D = dmat
N = dmat.shape[0]
a=np.tril(np.zeros([N,N])-1000000, k=-1)
# Get the lower triangle of the matrix.
C = np.triu(D)+a
# Mask the upper triangle.
C = np.ma.masked_array(C, C == -1000000)
# Set the diagonal to zero.
for i in range(N):
C[i, i] = 0
# Transformation matrix for rotating the heatmap.
A = np.array([(y, x) for x in range(N, -1, -1) for y in range(N + 1)])
t = np.array([[2**(-0.5), 2**(-0.5)], [-2**(-0.5), 2**(-0.5)]])
A = np.dot(A, t)
#t_ = np.array([[2**(-0.5), -2**(-0.5)], [2**(-0.5), 2**(-0.5)]])
# -1.0 correlation is blue, 0.0 is white, 1.0 is red.
cmap = plt.cm.Reds
#norm = mp.colors.BoundaryNorm(np.linspace(0, 10, 14), cmap.N)
# This MUST be before the call to pl.pcolormesh() to align properly.
ax.set_xticks([])
ax.set_yticks([])
X = A[:, 1].reshape(N + 1, N + 1)
Y = A[:, 0].reshape(N + 1, N + 1)
caxes = plt.pcolormesh(X, Y, np.flipud(C), axes=ax, cmap=cmap, rasterized=True)
x1s=[0]
if len(grid_pos)>0:
for i, grid in enumerate(grid_pos):
x0=(grid)/(2**(0.5))
x1=(grid)*(2**0.5)
y0=(grid)/(2**(0.5))
y1=0
ax.plot([x0, x1], [y0, y1],color='gray', linewidth=1)
ax.plot([(grid)*(2**0.5), (grid)*(2**0.5)+(N-grid)/(2**(0.5))], [0, (N-grid)/(2**(0.5))], color='gray', linewidth=1)
x1s.append(x1)
x1s.append(N*2**0.5)
if len(genes) >0:
leng=0
for i, g in enumerate(genes):
#ax.plot([(i+0.5)*(2**0.5), (i+0.5)*(2**0.5)], [-2**0.5, -(N/10)*2**0.5/2],color='b', linewidth=0.5)
ax.text((i+0.5)*(2**0.5), -(N/10)*2**0.5/2*1.01, g, rotation=90,ha='center', va='top', fontsize="small")
leng+=len(g)
leng=leng/len(genes)
else:
leng=0
spacing=(N/10)*leng*0.4
rect=Rectangle([0, -spacing],N*(2**0.5), spacing, color='whitesmoke' , alpha=1, linewidth=0)
ax.add_patch(rect)
rect=Rectangle([0, -(N/10)*2**0.5/2],N*(2**0.5), (N/10)*2**0.5/2, color='dimgray' , alpha=1, linewidth=0)
ax.add_patch(rect)
if len(grid_labels)>0:
for i in range(len(x1s)-1):
if i%2==1:
rect=Rectangle([x1s[i], -spacing],np.abs(x1s[i] - x1s[i+1]), spacing, color='silver' , alpha=0.3, linewidth=0)
ax.add_patch(rect)
else:
rect=Rectangle([x1s[i], -spacing],np.abs(x1s[i] - x1s[i+1]), spacing, color='lavender' , alpha=0.3, linewidth=0)
ax.add_patch(rect)
x=(x1s[i]+x1s[i+1])/2
ax.text(x,-(N/10)*2**0.5/4, grid_labels[i], rotation=90,ha='center', va='center', color="w")
cb = plt.colorbar(caxes, ax=ax, shrink=0.75)
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
ax.spines['bottom'].set_visible(False)
ax.spines['left'].set_visible(False)
if save!="":
if save.endswith(".pdf") or save.endswith(".png") or save.endswith(".svg"):
plt.savefig(save)
else:
plt.savefig(save+"_complexheatmap.pdf")
if show:
plt.show()
return {"axes": ax}
def complex_clustermap(df: pd.DataFrame,
variables: list=[],
dfcol: Optional[pd.DataFrame]=None,
row_colors: list=[],
col_colors: list=[],
row_plot: list=[],
col_plot: list=[],
row_scatter: list=[],
col_scatter: list=[],
row_bar: list=[],
col_bar: list=[],
ctitle: str="",
approx_clusternum: int=10,
approx_clusternum_col: int=3,
color_var: int=0,
merginalsum: bool=False,
show: bool=False,
method: str="ward",
return_col_cluster: bool=True,
ztranform: bool=True,
xticklabels: bool=True,
yticklabels: bool=False,
show_plot_labels: bool=False,
figsize: list=[],
title: str="",
save: str="",
heatmap_palette: str="coolwarm",
heatmap_col: list=[],
**kwargs):
"""
Drawing a clustered heatmap with merginal plots.
Parameters
----------
df : pandas DataFrame
variables : list
the column names of variables for heatmap
row_colors, col_colors: list, optional
the column names of categorical values to be plotted as color labels. for the col_colors to be plotted, dfcol options will be needed.
row_plot, col_plot : list, optional
The column names for the values to be plotted as lines.
row_scatter, col_scatter: list, optional
The column names for the values to be plotted as points.
row_bar, col_bar: list, optional
The column names for the values to be plotted as bars.
approx_clusternum : int, optional
The approximate number of row clusters to be created. Labeling the groups of leaves with different colors. The result of hierarchical clustering won't change.
approx_clusternum_col : int, optional
The approximate number of column clusters to be created. Labeling the groups of leaves with different colors. The result of hierarchical clustering won't change.
color_var : int, optional
The number of potential colors in dendrograms. If some clusters in the dendrogram share a same color (because the number of clusters is too many),
give this option may solve the problem.
merginalsum : bool, optional
Whether or not to draw bar plots for merginal distribution.
show : bool, optional
Whether or not to show the figure.
method : string, optional (default: "ward")
Method for hierarchical clustering. ["ward", "single",
return_col_cluster : string
The title for color values. If not set, "color_val" will be used.
ztranform: bool, optional (default: True)
xticklabels: bool, optional (default: True)
Whether or not to show xtick labels
yticklabels: bool, optional (default: False)
Whether or not to show ytick labels
show_plot_labels: bool, optional (default: False)
Whether or not to show plot labels.
figsize: list, optional
title: str, optional
save: str, optional
heatmap_col: list, optional
The same as the "variables" option. Will be deprecated.
Returns
-------
{"row_clusters":pd.DataFrame,"col_clusters":pd.DataFrame, "grid":g}: dict
Raises
------
Notes
-----
References
----------
See Also
--------
Examples
--------
"""#print(kwargs)
rnum, cnum=df.shape
if len(heatmap_col)==0 and len(variables)==0:
raise Exception("Please specify the variables option")
if len(heatmap_col)!=0:
variables=heatmap_col
cnum=len(variables)
if len(figsize)==0:
xsize=np.amin([cnum, 20])
xsize=np.amax([xsize, 5])
figsize=[xsize,10]
scatterpointsize=5
sns.set(font_scale=1)
if ztranform==True:
df[variables]=df[variables].apply(zscore)
if ctitle =="":
ctitle="zscore"
if len(col_plot)!=0 or len(col_scatter)!=0 or len(col_bar)!=0:
if dfcol==None:
raise Exception("if you want to plot along the x axis, you need provide dfcol option containing values to plot.")
cdata={"Cluster":[],"Index":[],"RGB":[]}
totalrowplot=0
if merginalsum==True:
totalrowplot+=1
totalrowplot+=len(row_plot)
totalrowplot+=len(row_colors)
totalrowplot+=len(row_scatter)
totalrowplot+=len(row_bar)
totalcolplot=0
if merginalsum==True:
totalcolplot+=1
totalcolplot+=len(col_plot)
totalcolplot+=len(col_colors)
totalcolplot+=len(col_scatter)
totalcolplot+=len(col_bar)
_row_color_legend={}
_col_color_legend={}
colormap_index=0
if totalrowplot + totalcolplot >0:
rowplotcount=0
colplotcount=0
_row_colors=[]
_row_colors_title=[]
_col_colors=[]
_col_colors_title=[]
if merginalsum:
_row_colors.append(np.ones([rnum, 4]))
_row_colors_title.append("Sum")
_col_colors.append(np.ones([cnum, 4]))
_col_colors_title.append("Sum")
#print(np.shape(_col_colors))
if len(row_colors)>0:
for k in row_colors:
u=np.array(natsorted(np.unique(df[k])))
_cmap=plt.get_cmap(colormap_list[colormap_index],u.shape[0])
lut={}
for _i, _u in enumerate(u):
lut[_u]=_cmap(_i)
_row_color_legend[k]=lut
colormap_index+=1
_row_colors.append([lut[cat] for cat in df[k]])
_row_colors_title.append(k)
if len(col_colors)>0:
for k in col_colors:
u=np.array(natsorted(np.unique(dfcol[k])))
_cmap=plt.get_cmap(colormap_list[colormap_index],u.shape[0])
lut={}
for _i, _u in enumerate(u):
lut[_u]=_cmap(_i)
_col_color_legend[k]=lut
colormap_index+=1
_col_colors.append([lut[cat] for cat in dfcol[k]])
_col_colors_title.append(k)
if len(row_plot)>0:
for k in row_plot:
_row_colors.append(np.ones([rnum, 4]))
_row_colors_title.append(k)
if len(col_plot)>0:
for k in col_plot:
_col_colors.append(np.ones([rnum, 4]))
_col_colors_title.append(k)
if len(row_scatter)>0:
for k in row_scatter:
_row_colors.append(np.ones([rnum, 4]))
_row_colors_title.append(k)
if len(col_scatter)>0:
for k in col_scatter:
_col_colors.append(np.ones([rnum, 4]))
_col_colors_title.append(k)
if len(row_bar)>0:
for k in row_bar:
_row_colors.append(np.ones([rnum, 4]))
_row_colors_title.append(k)
if len(col_bar)>0:
for k in col_bar:
_col_colors.append(np.ones([rnum, 4]))
_col_colors_title.append(k)
if len(_row_colors) >0 and len(_col_colors) >0:
#print(np.shape(_col_colors))
g=sns.clustermap(df[variables],col_colors=_col_colors,
row_colors=_row_colors,
method=method,xticklabels=xticklabels, yticklabels=yticklabels,
figsize=figsize,dendrogram_ratio=0.1,
cbar_kws=dict(cmap=heatmap_palette, orientation='horizontal'),
**kwargs)
g.ax_col_colors.invert_yaxis()
g.ax_row_colors.invert_xaxis()
elif len(_col_colors) >0:
g=sns.clustermap(df[variables],
col_colors=_col_colors,
method=method,
xticklabels=xticklabels,
yticklabels=yticklabels,
dendrogram_ratio=0.1,
cbar_kws=dict(cmap=heatmap_palette, orientation='horizontal'),
figsize=figsize,**kwargs)
g.ax_col_colors.invert_yaxis()
elif len(_row_colors) >0:
g=sns.clustermap(df[variables],
row_colors=_row_colors,
method=method,
xticklabels=xticklabels,
cmap=heatmap_palette,
yticklabels=yticklabels,
dendrogram_ratio=0.1,
figsize=figsize,
cbar_kws=dict(orientation='horizontal'),
**kwargs)
g.ax_row_colors.invert_xaxis()
#for spine in g.ax_cbar.spines:
# g.ax_cbar.spines[spine].set_linewidth(2)
rowplotcount=0
colplotcount=0
tickpos=0.9
row_labels=[]
col_labels=[]
row_ticks=[]
col_ticks=[]
if merginalsum:
mat=df[variables].to_numpy()
r=np.sum(mat, axis=1)
row_labels.append(0)
row_labels.append(np.amax(r))
row_ticks.append(rowplotcount)
row_ticks.append(rowplotcount+tickpos)
row_cluster=True
if "row_cluster" in kwargs:
row_cluster=kwargs["row_cluster"]
if row_cluster==True:
g.ax_row_colors.barh(np.arange(r.shape[0])+0.5, r[leaves_list(g.dendrogram_row.linkage)]/np.amax(r),height=1)
else:
g.ax_row_colors.barh(np.arange(r.shape[0])+0.5, r/np.amax(r),height=1)
#g.ax_row_colors.set_xticks([0,1],labels=[0,np.amax(r)])
c=np.sum(mat, axis=0)
#print(mat, c)
col_cluster=True
if "col_cluster" in kwargs:
col_cluster=kwargs["col_cluster"]
if col_cluster==True:
#print(leaves_list(g.dendrogram_col.linkage))
g.ax_col_colors.bar(np.arange(c.shape[0])+0.5,c[leaves_list(g.dendrogram_col.linkage)]/np.amax(c),width=1)
else:
g.ax_col_colors.bar(np.arange(c.shape[0])+0.5,c/np.amax(c),width=1)
#g.ax_col_colors.set_yticks([0,1],labels=[0,np.amax(c)])
col_labels.append(0)
col_labels.append(np.amax(c))
col_ticks.append(colplotcount)
col_ticks.append(colplotcount+tickpos)
rowplotcount=1
colplotcount=1
rowplotcount+=len(row_colors)
colplotcount+=len(col_colors)
if len(row_plot)>0:
row_cluster=True
if "row_cluster" in kwargs:
row_cluster=kwargs["row_cluster"]
for i, lname in enumerate(row_plot):
r=np.array(df[lname])
row_labels.append(np.amin(r))
row_labels.append(np.amax(r))
row_ticks.append(rowplotcount)
row_ticks.append(rowplotcount+tickpos)
r=r-np.amin(r)
r=r/np.amax(r)
r=0.9*r
if row_cluster==True:
tmpindx=leaves_list(g.dendrogram_row.linkage)
r=r[tmpindx]
g.ax_row_colors.plot(r+rowplotcount, np.arange(r.shape[0])+0.5)
else:
g.ax_row_colors.plot(r+rowplotcount, np.arange(r.shape[0])+0.5)
rowplotcount+=1
if len(col_plot)>0:
col_cluster=True
if "col_cluster" in kwargs:
col_cluster=kwargs["col_cluster"]
for i, lname in enumerate(col_plot):
r=np.array(dfcol[lname])
col_labels.append(np.amin(r))
col_labels.append(np.amax(r))
col_ticks.append(colplotcount)
col_ticks.append(colplotcount+tickpos)
r=r-np.amin(r)
r=r/np.amax(r)
r=0.9*r
if col_cluster==True:
g.ax_col_colors.plot(np.arange(r.shape[0])+0.5,r[leaves_list(g.dendrogram_col.linkage)]+colplotcount)
else:
g.ax_col_colors.plot(np.arange(r.shape[0])+0.5,r+colplotcount)
colplotcount+=1
if len(row_scatter)>0:
row_cluster=True
if "row_cluster" in kwargs:
row_cluster=kwargs["row_cluster"]
for i, lname in enumerate(row_scatter):
r=np.array(df[lname])
row_labels.append(np.amin(r))
row_labels.append(np.amax(r))
row_ticks.append(rowplotcount)
row_ticks.append(rowplotcount+tickpos)
r=r-np.amin(r)
r=r/np.amax(r)
r=0.9*r
if row_cluster==True:
tmpindx=leaves_list(g.dendrogram_row.linkage)
r=r[tmpindx]
g.ax_row_colors.scatter(r+rowplotcount, np.arange(r.shape[0])+0.5,s=scatterpointsize)
else:
g.ax_row_colors.scatter(r+rowplotcount, np.arange(r.shape[0])+0.5,s=scatterpointsize)
rowplotcount+=1
if len(col_scatter)>0:
col_cluster=True
if "col_cluster" in kwargs:
col_cluster=kwargs["col_cluster"]
for i, lname in enumerate(col_scatter):
r=np.array(dfcol[lname])
col_labels.append(np.amin(r))
col_labels.append(np.amax(r))
col_ticks.append(colplotcount)
col_ticks.append(colplotcount+tickpos)
r=r-np.amin(r)
r=r/np.amax(r)
r=0.9*r
if col_cluster==True:
g.ax_col_colors.bar(np.arange(r.shape[0])+0.5,r[leaves_list(g.dendrogram_col.linkage)]+colplotcount,s=scatterpointsize)
else:
g.ax_col_colors.bar(np.arange(r.shape[0])+0.5,r+colplotcount,s=scatterpointsize)
colplotcount+=1
if len(row_bar)>0:
row_cluster=True
if "row_cluster" in kwargs:
row_cluster=kwargs["row_cluster"]
for i, lname in enumerate(row_bar):
r=np.array(df[lname])
row_labels.append(np.amin(r))
row_labels.append(np.amax(r))
row_ticks.append(rowplotcount)
row_ticks.append(rowplotcount+tickpos)
r=r-np.amin(r)
r=r/np.amax(r)
r=0.9*r
if row_cluster==True:
g.ax_row_colors.barh(y=np.arange(r.shape[0])+0.5, width=r[leaves_list(g.dendrogram_row.linkage)],left=[rowplotcount]*r.shape[0])
else:
g.ax_row_colors.barh(r, np.arange(r.shape[0])+0.5,left=[rowplotcount]*r.shape[0])
rowplotcount+=1
if len(col_bar)>0:
col_cluster=True
if "col_cluster" in kwargs:
col_cluster=kwargs["col_cluster"]
for i, lname in enumerate(col_bar):
r=np.array(dfcol[lname])
col_labels.append(np.amin(r))
col_labels.append(np.amax(r))
col_ticks.append(colplotcount)
col_ticks.append(colplotcount+tickpos)
if col_cluster==True:
g.ax_col_colors.scatter(np.arange(r.shape[0])+0.5,r[leaves_list(g.dendrogram_col.linkage)]/(np.amax(r)*1.1)+colplotcount)
else:
g.ax_col_colors.scatter(np.arange(r.shape[0])+0.5,r/(np.amax(r)*1.1)+colplotcount)
colplotcount+=1
if g.ax_row_colors!=None:
g.ax_row_colors.set_xticks(np.arange(len(_row_colors_title))+0.5)
g.ax_row_colors.set_xticklabels(_row_colors_title, rotation=90)
if g.ax_col_colors!=None:
colax_otherside = g.ax_col_colors.twinx()
colax_otherside.set_yticks(0.5*(np.arange(len(_col_colors_title))+0.5),labels=_col_colors_title)
colax_otherside.grid(False)
if g.ax_col_dendrogram!=None:
col = g.ax_col_dendrogram.get_position()
g.ax_col_dendrogram.set_position([col.x0, col.y0, col.width*0.5, col.height*0.5])
if show_plot_labels==True:
rowax_otherside = g.ax_row_colors.twiny()
rowax_otherside.invert_xaxis()
rowax_otherside.grid(False)
rowax_otherside.set_xticks(row_ticks, labels=np.round(row_labels,2), rotation=90, fontsize=8)
col_ticks=np.array(col_ticks)
g.ax_col_colors.set_yticks(col_ticks, labels=np.round(col_labels,2), fontsize=8)
legend_num=0
for _title, colorlut in _row_color_legend.items():
legendhandles=[]
for label, color in colorlut.items():
legendhandles.append(Line2D([0], [0], color=color,linewidth=5, label=label))
legend1=g.ax_heatmap.legend(handles=legendhandles,
loc="upper left",
title=_title,
bbox_to_anchor=(1.15, 1-0.2*legend_num))
g.ax_heatmap.add_artist(legend1)
legend_num+=1
for _title, colorlut in _col_color_legend.items():
legendhandles=[]
for label, color in colorlut.items():
legendhandles.append(Line2D([0], [0], color=color,linewidth=5, label=label))
legend1=g.ax_heatmap.legend(handles=legendhandles,
loc="upper left",
title=_title,
bbox_to_anchor=(1.15, 1-0.2*legend_num))
g.ax_heatmap.add_artist(legend1)
legend_num+=1
else:
g=sns.clustermap(df,method=method,
cbar_kws=dict(orientation='horizontal'),
xticklabels=xticklabels,
cmap=heatmap_palette,
yticklabels=yticklabels,
dendrogram_ratio=0.1,
figsize=figsize,
**kwargs)
if color_var>0:
cmap = cm.tab20c(np.linspace(0, 1, color_var))
else:
cmap = cm.tab20c(np.linspace(0, 1, approx_clusternum+5))
hierarchy.set_link_color_palette([mpl.colors.rgb2hex(rgb[:3]) for rgb in cmap])
"""coloring the row dendrogram based on branch numbers crossed with the threshold"""
if g.dendrogram_row != None:
t=_dendrogram_threshold(g.dendrogram_row.dendrogram,approx_clusternum)
den=hierarchy.dendrogram(g.dendrogram_row.linkage,
labels = g.data.index,
color_threshold=t,ax=g.ax_row_dendrogram,
orientation="left")
g.ax_row_dendrogram.invert_yaxis()
clusters = _get_cluster_classes(den)
keys=list(clusters.keys())
ckeys={}
i=1
for k in keys:
if k=="C0":
ckeys[k]="C0"
else:
ckeys[k]="C"+str(i)
i+=1
for c, v in clusters.items():
_c=ckeys[c]
for _v in v:
cdata["Cluster"].append(_c)
cdata["Index"].append(_v)
cdata["RGB"].append(matplotlib.colors.to_rgb(c))
"""Setting the row dendrogram ends here"""
"""coloring the col dendrogram based on branch numbers crossed with the threshold"""
col_cdata={"Cluster":[],"Index":[],"RGB":[]}
if g.dendrogram_col != None:
t=_dendrogram_threshold(g.dendrogram_col.dendrogram,approx_clusternum_col)
den=hierarchy.dendrogram(g.dendrogram_col.linkage,
labels = g.data.columns,
color_threshold=t,ax=g.ax_col_dendrogram,
orientation="top")
#g.ax_col_dendrogram.invert_yaxis()
col_clusters = _get_cluster_classes(den)
col_cdata={"Cluster":[],"Index":[],"RGB":[]}
col_keys=list(col_clusters.keys())
col_ckeys={}
i=1
for k in col_keys:
if k=="C0":
col_ckeys[k]="C0"
else:
col_ckeys[k]="C"+str(i)
i+=1
for c, v in col_clusters.items():
_c=col_ckeys[c]
for _v in v:
col_cdata["Cluster"].append(_c)
col_cdata["Index"].append(_v)
col_cdata["RGB"].append(matplotlib.colors.to_rgb(c))
"""Setting the col dendrogram ends here"""
g.fig.subplots_adjust(bottom=0.175, right=0.70)
x0, _y0, _w, _h = g.cbar_pos
g.ax_cbar.set_title(ctitle)
#g.ax_cbar.tick_params(axis='x', length=5)
g.ax_cbar.set_position(pos=[0.80, 0.1, 0.1, 0.02],
which="both")
if title !="":
g.fig.suptitle(title, va="bottom")
plt.setp(g.ax_heatmap.xaxis.get_majorticklabels(), rotation=90)
_save(save, "complex_clustermap")
if show:
plt.show()
return {"data":g.data2d,"row_clusters":pd.DataFrame(cdata),"col_clusters":pd.DataFrame(col_cdata), "grid":g}
def dotplot(df: pd.DataFrame,
y: str="",
x: str="",
dfc=pd.DataFrame(),
scaling: float=10,
color_val: str="",
size_val: str="",
highlight: str="",
color_title: str="",
size_title: str="",
figsize: list=[],
save: str="",
threshold: float=-np.log10(0.05),
row_clustering: bool=True,
xtickrotation: float=90,
column_order: list=[],
colorpalette="coolwarm",
show: bool=False,
title: str="",
row: str="",
col: str="",
) -> Dict[str, plt.Axes]:
"""
Drawing a dotplot that can represent two different variables as dot sizes and colors on a regular grid.
This function is assumed to plot GO enrichment analysis with multiple gene sets.
Parameters
----------
df : pandas DataFrame
dataframe containing two categories and corresponding values (such as p values and odds ratio).
e.g.:
Cluster Condensate pval odds FDR
54 C1 Cajal body -0.000000 0.000000 -0.000000
55 C1 *DNA repair focus -0.000000 0.000000 -0.000000
56 C1 *DNA replication condensate -0.000000 0.000000 -0.000000
57 C1 P-body -0.000000 0.000000 -0.000000
58 C1 PML body -0.000000 0.000000 -0.000000
row: string
the column name of a category that is going to be placed in the row of the dotplot
col: string
the column name of a category that is going to be placed in the column of the dotplot
color_val : string
The column name for the values represented as dot colors.
size_val : string
The column name for the values represented as dot sizes.
scaling: float
The scale of dots. If resulting dots are too large (or small), you can reduce (or increase) dot sizes by adjusting this value.
highlight : string
A dictionary to set color labels to leaves. The key is the name of the color label.
The value is the list of RGB color codes, each corresponds to the color of a leaf.
e.g., {"color1":[[1,0,0,1], ....]}
size_title : string
The title for size values. If not set, "size_val" will be used.
color_title : string
The title for color values. If not set, "color_val" will be used.
show : bool
Whether or not to show the figure.
Returns
-------
axes: dict {"axes1":ax1,"axes2":ax2,"axes3":ax3}
Raises
------
Notes
-----
References
----------
See Also
--------
Examples
--------
"""
if x !="" and y!="":
row=y
col=x
if size_val!="":
_df=df.pivot_table(index=col,columns=row,values=size_val)
if len(column_order)>0:
_df=_df.loc[column_order]
else:
_df=_df.sort_index(axis=0,key=natsort_keygen())
_df=_df.fillna(0)
if color_val!="":
dfc=df.pivot_table(index=col,columns=row,values=color_val)
if len(column_order)>0:
dfc=dfc.loc[column_order]
else:
dfc=dfc.sort_index(axis=0,key=natsort_keygen())
dfc=dfc.fillna(0)
if highlight !="":
dfh=df.pivot_table(index=col,columns=row,values=highlight)
if len(column_order)>0:
dfh=dfh.loc[column_order]
else:
dfh=dfh.sort_index(axis=0,key=natsort_keygen())
dfh=dfh.fillna(0)
if row_clustering==True:
Y = fcl.linkage(_df.to_numpy().T, method='ward', metric='euclidean')
Z1 = sch.dendrogram(Y,no_plot=True)
idx1 = Z1['leaves']
reorder=_df.columns[idx1]
_df=_df[reorder]
if color_val!="":
dfc=dfc[reorder]
if highlight !="":
dfh=dfh[reorder]
_x=_df.index
_y=_df.columns
mat=_df.to_numpy()
minsize=np.amin(mat)
maxsize=np.amax(mat)
else:
_x=df.index
_y=df.columns
mat=df.to_numpy()
minsize=np.amin(mat)
maxsize=np.amax(mat)
#if minsize==0:
#mat=mat+maxsize*0.01
#minsize=np.amin(mat)
maxsize=np.round(maxsize)
middle0=np.round((minsize+maxsize)/3)
middle1=np.round(2*(minsize+maxsize)/3)
#plt.rcParams["figure.figsize"] = [7.50, 3.50]
#plt.rcParams["figure.autolayout"] = True
#x = np.arange(len(_x))
#y = np.arange(len(_y))
#X, Y = np.meshgrid(x, y)
xy=[ [i,j] for i in range(len(_x)) for j in range(len(_y))]
#num = 1000
sizes = [mat[i,j]*scaling for i in range(len(_x)) for j in range(len(_y))]
edge_colors=[]
if highlight !="":
hmat=dfh.to_numpy()
hvals = [hmat[i,j]*scaling for i in range(len(_x)) for j in range(len(_y))]
for s in hvals:
if s>=threshold*scaling:
edge_colors.append("magenta")
else:
edge_colors.append("gray")
else:
for s in sizes:
if s>=threshold*scaling:
edge_colors.append("magenta")
else:
edge_colors.append("gray")
if len(dfc) !=0:
viridis = cm.get_cmap(colorpalette, 12)
cmat=dfc.to_numpy()
cmat[cmat==np.inf]=0
_cmat=cmat/np.amax(cmat)
_colors = [viridis(_cmat[i,j]) for i in range(len(_x)) for j in range(len(_y))]
else:
_colors = [[0,1,0,1] for i in range(len(_x)) for j in range(len(_y))]
#print(sizes)
#xy = 10 * np.random.random((num, 2))
#xy=XY
#patches = [plt.Circle(center, size) for center, size in zip(xy, sizes)]
#fig, ax = plt.subplots(ncols=2, gridspec_kw={'width_ratios': [8, 2]})
if len(figsize)==0:
figsize=[mat.shape[0]*0.5+2,mat.shape[1]*0.5+1]
fig = plt.figure(figsize=figsize)
#fig.set_figheight(6)
#fig.set_figwidth(6)
ax1 = plt.subplot2grid(shape=(10, 6), loc=(0, 0), colspan=4, rowspan=10)
ax2 = plt.subplot2grid(shape=(10, 6), loc=(1, 4), colspan=2, rowspan=4)
ax3 = plt.subplot2grid(shape=(10, 6), loc=(6, 4), colspan=2, rowspan=1)
collection = mc.CircleCollection(sizes,
edgecolors=edge_colors,
offsets=xy,
transOffset=ax1.transData,
facecolors=_colors,
linewidths=2)
ax1.add_collection(collection)
ax1.margins(0.1)
ax1.set_xlim(-0.5,len(_x)-0.5)
ax1.set_xticks(np.arange(len(_x)))
ax1.set_xticklabels(_x,rotation=xtickrotation)
ax1.set_yticks(np.arange(len(_y)))
ax1.set_yticklabels(_y, rotation=0)
if color_title=="":
color_title=color_val
if len(dfc) !=0:
norm = mpl.colors.Normalize(vmin=np.min(cmat), vmax=np.amax(cmat))
cb1 = mpl.colorbar.ColorbarBase(ax3, cmap=viridis,
norm=norm,
orientation='horizontal')
cb1.set_label(color_title)
#ax[1]=fig.add_axes([1,0.3,0.1,1])
lxy=[[0.5, i*0.5] for i in range(3)]
collection2 = mc.CircleCollection([middle0*scaling,middle1*scaling, maxsize*scaling],
offsets=lxy,
transOffset=ax2.transData,
facecolors='lightgray',
edgecolors="gray")
ax2.add_collection(collection2)
ax2.axis('off')
ax2.margins(0.3)
for text, (x, y) in zip([middle0,middle1, maxsize], lxy):
ax2.text(x+0.01, y,str(text), ha="left",va="center",color="black" )
if size_title=="":
size_title=size_val
ax2.text(0.5,-0.5, size_title,va="center",ha="center")
#ax[1].set_yticks(np.arange(3))
#ax[1].set_yticklabels([minsize,middle, maxsize], rotation=0)
#plt.tight_layout()
plt.subplots_adjust(left=0.3,bottom=0.2)
#plt.tight_layout()
if title !="":
fig.suptitle(title)
_save(save, "dotplot")
if show==True:
plt.show()
return {"axes1":ax1,"axes2":ax2,"axes3":ax3}
def violinplot(df: pd.DataFrame,
x: str,
y: str,
pairs: list=[],
test: str="ttest_ind",
alternative: str="two-sided",
significance: str="numeric",
significance_ranges: Dict[str, float]={"*":-np.log10(0.05),"**":4,"***":10},
swarm: bool=False,
xorder: list=[],
equal_var: bool=False,
yunit: str="",
title: str="",
save: str="",
ax: Optional[plt.Axes]=None,**kwargs):
"""
Draw a boxplot with a statistical test
Parameters
----------
df : pandas DataFrame
x,y: str
names of variables in data
pairs: list, optional
Category pairs for the statistical test.
Examples: [["Adelie","Chinstrap" ],
["Gentoo","Chinstrap" ],
["Adelie","Gentoo" ]]
test: str, optional
Method name for the statistical test. Defalt: ttest_ind
Available methods: ["ttest_ind",
"ttest_rel",
"kruskal",
"mannwhitneyu",
"wilcoxon",
"brunnermunzel",
"median_test"]
alternative: str ['two-sided', 'less', 'greater'], optional
Defines the alternative hypothesis. Defalt: "two-sided"
show : bool, optional
Whether or not to show the figure.
significance: str ['numeric', 'symbol'], optional
How to show the significance. 'numeric' will show -log10(p values) in the plot and
'symbol' will represent significance as asterisks.
significance_ranges: dict, optional
thresholds of -log10(p values) that each asterisk number represents. Ignored when significance="numeric".
example: {"*":-np.log10(0.05),"**":4,"***":10}
swarm: bool, optional
Whether or not to superpose a swarm plot. Not recommended if the sample size is too large.
xorder: list, optional
The order of x axis labels
equal_var: bool, optional
Related to ttest_ind method. The default is True, which will produce a p value equal to t-test in R.
kwargs: any options accepted by scipy statistical test functions
Returns
-------
dict("p values":pvalues,"axes":ax)
Raises
------
Notes
-----
References
----------
See Also
--------
Examples
--------
"""
tests=["ttest_ind","ttest_rel","kruskal","mannwhitneyu","wilcoxon","brunnermunzel","median_test"]
if not test in tests:
raise Exception("Available tests are "+", ".join(tests))
import scipy.stats as stats
if len(xorder)==0:
xorder=sorted(list(set(df[x])))
pvals=[]
for p1,p2 in pairs:
statstest=getattr(stats, test)
if test=="wilcoxon" or test=="ttest_rel":
_, pval,_=statstest(df[y][df[x]==p1],df[y][df[x]==p2],alternative=alternative,**kwargs)
elif test=="median_test":
_, pval,_,_=statstest(df[y][df[x]==p1],df[y][df[x]==p2],alternative=alternative,**kwargs)
elif test=="ttest_ind":
_, pval=statstest(df[y][df[x]==p1],df[y][df[x]==p2],alternative=alternative,equal_var=equal_var,**kwargs)
else:
_, pval=statstest(df[y][df[x]==p1],df[y][df[x]==p2],alternative=alternative,**kwargs)
p1ind=xorder.index(p1)
p2ind=xorder.index(p2)
if pval==0:
pval=np.inf
else:
pval=-np.log10(pval)
pvals.append([np.abs(p2ind-p1ind), np.amin([p2ind, p1ind]),np.amax([p2ind, p1ind]), pval])
pvals = sorted(pvals, key = lambda x: (x[0], x[1]))
fig, ax=plt.subplots()
sns.violinplot(data=df, x=x,y=y,inner="quartile")
if swarm==True:
sns.swarmplot(data=df, x=x,y=y,color="black",alpha=0.5)
ymax=np.amax(df[y])
newpvals={}
for i, pval in enumerate(pvals):
plt.plot([pval[1],pval[2]], [ymax*(1.05+i*0.05),ymax*(1.05+i*0.05)], color="black")
p=np.round(pval[-1],2)
newpvals[xorder[pval[1]]+"_"+xorder[pval[2]]]=p
if significance=="numeric":
annotate="-log10(p)="+str(p)
elif significance=="symbol":
keys=sorted(significance_ranges.keys())
annotate="NA"
for j in range(len(keys)):
if j==0:
if p <= significance_ranges[keys[j]]:
annotate=""
break
else:
if significance_ranges[keys[j-1]] < p <=significance_ranges[keys[j]]:
annotate=keys[i]
break
if annotate=="NA":
annotate=keys[-1]
plt.text((pval[1]+pval[2])/2, ymax*(1.055+i*0.05), annotate)
if significance=="symbol":
ax.annotate("\n".join(["{}: p < {:.2E}".format(k, 10**(-significance_ranges[k])) for k in keys]),
xy=(0.9,0.9), xycoords='axes fraction',
textcoords='offset points',
size=12,
bbox=dict(boxstyle="round", fc=(0.9, 0.9, 0.9), ec="none"))
plt.subplots_adjust(right=0.850)
if yunit!="":
ax.text(0, 1, "({})".format(yunit), transform=ax.transAxes, ha="right")
_save(save, "violin")
return {"p values":newpvals,"axes":ax}
if __name__=="__main__":
passFunctions
def complex_clustermap(df: pandas.core.frame.DataFrame, variables: list = [], dfcol: Optional[pandas.core.frame.DataFrame] = None, row_colors: list = [], col_colors: list = [], row_plot: list = [], col_plot: list = [], row_scatter: list = [], col_scatter: list = [], row_bar: list = [], col_bar: list = [], ctitle: str = '', approx_clusternum: int = 10, approx_clusternum_col: int = 3, color_var: int = 0, merginalsum: bool = False, show: bool = False, method: str = 'ward', return_col_cluster: bool = True, ztranform: bool = True, xticklabels: bool = True, yticklabels: bool = False, show_plot_labels: bool = False, figsize: list = [], title: str = '', save: str = '', heatmap_palette: str = 'coolwarm', heatmap_col: list = [], **kwargs)-
Drawing a clustered heatmap with merginal plots.
Parameters
df:pandas DataFramevariables:list- the column names of variables for heatmap
row_colors,col_colors:list, optional- the column names of categorical values to be plotted as color labels. for the col_colors to be plotted, dfcol options will be needed.
row_plot,col_plot:list, optional- The column names for the values to be plotted as lines.
row_scatter,col_scatter:list, optional- The column names for the values to be plotted as points.
row_bar,col_bar:list, optional- The column names for the values to be plotted as bars.
approx_clusternum:int, optional- The approximate number of row clusters to be created. Labeling the groups of leaves with different colors. The result of hierarchical clustering won't change.
approx_clusternum_col:int, optional- The approximate number of column clusters to be created. Labeling the groups of leaves with different colors. The result of hierarchical clustering won't change.
color_var:int, optional- The number of potential colors in dendrograms. If some clusters in the dendrogram share a same color (because the number of clusters is too many), give this option may solve the problem.
merginalsum:bool, optional- Whether or not to draw bar plots for merginal distribution.
show:bool, optional- Whether or not to show the figure.
method:string, optional(default: "ward")- Method for hierarchical clustering. ["ward", "single",
return_col_cluster:string- The title for color values. If not set, "color_val" will be used.
ztranform:bool, optional(default: True)xticklabels:bool, optional(default: True)- Whether or not to show xtick labels
yticklabels:bool, optional(default: False)- Whether or not to show ytick labels
show_plot_labels:bool, optional(default: False)- Whether or not to show plot labels.
figsize:list, optionaltitle:str, optionalsave:str, optionalheatmap_col:list, optional- The same as the "variables" option. Will be deprecated.
Returns
{"row_clusters":pd.DataFrame,"col_clusters":pd.DataFrame, "grid":g}: dictRaises
Notes
References
See Also
Examples
Expand source code
def complex_clustermap(df: pd.DataFrame, variables: list=[], dfcol: Optional[pd.DataFrame]=None, row_colors: list=[], col_colors: list=[], row_plot: list=[], col_plot: list=[], row_scatter: list=[], col_scatter: list=[], row_bar: list=[], col_bar: list=[], ctitle: str="", approx_clusternum: int=10, approx_clusternum_col: int=3, color_var: int=0, merginalsum: bool=False, show: bool=False, method: str="ward", return_col_cluster: bool=True, ztranform: bool=True, xticklabels: bool=True, yticklabels: bool=False, show_plot_labels: bool=False, figsize: list=[], title: str="", save: str="", heatmap_palette: str="coolwarm", heatmap_col: list=[], **kwargs): """ Drawing a clustered heatmap with merginal plots. Parameters ---------- df : pandas DataFrame variables : list the column names of variables for heatmap row_colors, col_colors: list, optional the column names of categorical values to be plotted as color labels. for the col_colors to be plotted, dfcol options will be needed. row_plot, col_plot : list, optional The column names for the values to be plotted as lines. row_scatter, col_scatter: list, optional The column names for the values to be plotted as points. row_bar, col_bar: list, optional The column names for the values to be plotted as bars. approx_clusternum : int, optional The approximate number of row clusters to be created. Labeling the groups of leaves with different colors. The result of hierarchical clustering won't change. approx_clusternum_col : int, optional The approximate number of column clusters to be created. Labeling the groups of leaves with different colors. The result of hierarchical clustering won't change. color_var : int, optional The number of potential colors in dendrograms. If some clusters in the dendrogram share a same color (because the number of clusters is too many), give this option may solve the problem. merginalsum : bool, optional Whether or not to draw bar plots for merginal distribution. show : bool, optional Whether or not to show the figure. method : string, optional (default: "ward") Method for hierarchical clustering. ["ward", "single", return_col_cluster : string The title for color values. If not set, "color_val" will be used. ztranform: bool, optional (default: True) xticklabels: bool, optional (default: True) Whether or not to show xtick labels yticklabels: bool, optional (default: False) Whether or not to show ytick labels show_plot_labels: bool, optional (default: False) Whether or not to show plot labels. figsize: list, optional title: str, optional save: str, optional heatmap_col: list, optional The same as the "variables" option. Will be deprecated. Returns ------- {"row_clusters":pd.DataFrame,"col_clusters":pd.DataFrame, "grid":g}: dict Raises ------ Notes ----- References ---------- See Also -------- Examples -------- """#print(kwargs) rnum, cnum=df.shape if len(heatmap_col)==0 and len(variables)==0: raise Exception("Please specify the variables option") if len(heatmap_col)!=0: variables=heatmap_col cnum=len(variables) if len(figsize)==0: xsize=np.amin([cnum, 20]) xsize=np.amax([xsize, 5]) figsize=[xsize,10] scatterpointsize=5 sns.set(font_scale=1) if ztranform==True: df[variables]=df[variables].apply(zscore) if ctitle =="": ctitle="zscore" if len(col_plot)!=0 or len(col_scatter)!=0 or len(col_bar)!=0: if dfcol==None: raise Exception("if you want to plot along the x axis, you need provide dfcol option containing values to plot.") cdata={"Cluster":[],"Index":[],"RGB":[]} totalrowplot=0 if merginalsum==True: totalrowplot+=1 totalrowplot+=len(row_plot) totalrowplot+=len(row_colors) totalrowplot+=len(row_scatter) totalrowplot+=len(row_bar) totalcolplot=0 if merginalsum==True: totalcolplot+=1 totalcolplot+=len(col_plot) totalcolplot+=len(col_colors) totalcolplot+=len(col_scatter) totalcolplot+=len(col_bar) _row_color_legend={} _col_color_legend={} colormap_index=0 if totalrowplot + totalcolplot >0: rowplotcount=0 colplotcount=0 _row_colors=[] _row_colors_title=[] _col_colors=[] _col_colors_title=[] if merginalsum: _row_colors.append(np.ones([rnum, 4])) _row_colors_title.append("Sum") _col_colors.append(np.ones([cnum, 4])) _col_colors_title.append("Sum") #print(np.shape(_col_colors)) if len(row_colors)>0: for k in row_colors: u=np.array(natsorted(np.unique(df[k]))) _cmap=plt.get_cmap(colormap_list[colormap_index],u.shape[0]) lut={} for _i, _u in enumerate(u): lut[_u]=_cmap(_i) _row_color_legend[k]=lut colormap_index+=1 _row_colors.append([lut[cat] for cat in df[k]]) _row_colors_title.append(k) if len(col_colors)>0: for k in col_colors: u=np.array(natsorted(np.unique(dfcol[k]))) _cmap=plt.get_cmap(colormap_list[colormap_index],u.shape[0]) lut={} for _i, _u in enumerate(u): lut[_u]=_cmap(_i) _col_color_legend[k]=lut colormap_index+=1 _col_colors.append([lut[cat] for cat in dfcol[k]]) _col_colors_title.append(k) if len(row_plot)>0: for k in row_plot: _row_colors.append(np.ones([rnum, 4])) _row_colors_title.append(k) if len(col_plot)>0: for k in col_plot: _col_colors.append(np.ones([rnum, 4])) _col_colors_title.append(k) if len(row_scatter)>0: for k in row_scatter: _row_colors.append(np.ones([rnum, 4])) _row_colors_title.append(k) if len(col_scatter)>0: for k in col_scatter: _col_colors.append(np.ones([rnum, 4])) _col_colors_title.append(k) if len(row_bar)>0: for k in row_bar: _row_colors.append(np.ones([rnum, 4])) _row_colors_title.append(k) if len(col_bar)>0: for k in col_bar: _col_colors.append(np.ones([rnum, 4])) _col_colors_title.append(k) if len(_row_colors) >0 and len(_col_colors) >0: #print(np.shape(_col_colors)) g=sns.clustermap(df[variables],col_colors=_col_colors, row_colors=_row_colors, method=method,xticklabels=xticklabels, yticklabels=yticklabels, figsize=figsize,dendrogram_ratio=0.1, cbar_kws=dict(cmap=heatmap_palette, orientation='horizontal'), **kwargs) g.ax_col_colors.invert_yaxis() g.ax_row_colors.invert_xaxis() elif len(_col_colors) >0: g=sns.clustermap(df[variables], col_colors=_col_colors, method=method, xticklabels=xticklabels, yticklabels=yticklabels, dendrogram_ratio=0.1, cbar_kws=dict(cmap=heatmap_palette, orientation='horizontal'), figsize=figsize,**kwargs) g.ax_col_colors.invert_yaxis() elif len(_row_colors) >0: g=sns.clustermap(df[variables], row_colors=_row_colors, method=method, xticklabels=xticklabels, cmap=heatmap_palette, yticklabels=yticklabels, dendrogram_ratio=0.1, figsize=figsize, cbar_kws=dict(orientation='horizontal'), **kwargs) g.ax_row_colors.invert_xaxis() #for spine in g.ax_cbar.spines: # g.ax_cbar.spines[spine].set_linewidth(2) rowplotcount=0 colplotcount=0 tickpos=0.9 row_labels=[] col_labels=[] row_ticks=[] col_ticks=[] if merginalsum: mat=df[variables].to_numpy() r=np.sum(mat, axis=1) row_labels.append(0) row_labels.append(np.amax(r)) row_ticks.append(rowplotcount) row_ticks.append(rowplotcount+tickpos) row_cluster=True if "row_cluster" in kwargs: row_cluster=kwargs["row_cluster"] if row_cluster==True: g.ax_row_colors.barh(np.arange(r.shape[0])+0.5, r[leaves_list(g.dendrogram_row.linkage)]/np.amax(r),height=1) else: g.ax_row_colors.barh(np.arange(r.shape[0])+0.5, r/np.amax(r),height=1) #g.ax_row_colors.set_xticks([0,1],labels=[0,np.amax(r)]) c=np.sum(mat, axis=0) #print(mat, c) col_cluster=True if "col_cluster" in kwargs: col_cluster=kwargs["col_cluster"] if col_cluster==True: #print(leaves_list(g.dendrogram_col.linkage)) g.ax_col_colors.bar(np.arange(c.shape[0])+0.5,c[leaves_list(g.dendrogram_col.linkage)]/np.amax(c),width=1) else: g.ax_col_colors.bar(np.arange(c.shape[0])+0.5,c/np.amax(c),width=1) #g.ax_col_colors.set_yticks([0,1],labels=[0,np.amax(c)]) col_labels.append(0) col_labels.append(np.amax(c)) col_ticks.append(colplotcount) col_ticks.append(colplotcount+tickpos) rowplotcount=1 colplotcount=1 rowplotcount+=len(row_colors) colplotcount+=len(col_colors) if len(row_plot)>0: row_cluster=True if "row_cluster" in kwargs: row_cluster=kwargs["row_cluster"] for i, lname in enumerate(row_plot): r=np.array(df[lname]) row_labels.append(np.amin(r)) row_labels.append(np.amax(r)) row_ticks.append(rowplotcount) row_ticks.append(rowplotcount+tickpos) r=r-np.amin(r) r=r/np.amax(r) r=0.9*r if row_cluster==True: tmpindx=leaves_list(g.dendrogram_row.linkage) r=r[tmpindx] g.ax_row_colors.plot(r+rowplotcount, np.arange(r.shape[0])+0.5) else: g.ax_row_colors.plot(r+rowplotcount, np.arange(r.shape[0])+0.5) rowplotcount+=1 if len(col_plot)>0: col_cluster=True if "col_cluster" in kwargs: col_cluster=kwargs["col_cluster"] for i, lname in enumerate(col_plot): r=np.array(dfcol[lname]) col_labels.append(np.amin(r)) col_labels.append(np.amax(r)) col_ticks.append(colplotcount) col_ticks.append(colplotcount+tickpos) r=r-np.amin(r) r=r/np.amax(r) r=0.9*r if col_cluster==True: g.ax_col_colors.plot(np.arange(r.shape[0])+0.5,r[leaves_list(g.dendrogram_col.linkage)]+colplotcount) else: g.ax_col_colors.plot(np.arange(r.shape[0])+0.5,r+colplotcount) colplotcount+=1 if len(row_scatter)>0: row_cluster=True if "row_cluster" in kwargs: row_cluster=kwargs["row_cluster"] for i, lname in enumerate(row_scatter): r=np.array(df[lname]) row_labels.append(np.amin(r)) row_labels.append(np.amax(r)) row_ticks.append(rowplotcount) row_ticks.append(rowplotcount+tickpos) r=r-np.amin(r) r=r/np.amax(r) r=0.9*r if row_cluster==True: tmpindx=leaves_list(g.dendrogram_row.linkage) r=r[tmpindx] g.ax_row_colors.scatter(r+rowplotcount, np.arange(r.shape[0])+0.5,s=scatterpointsize) else: g.ax_row_colors.scatter(r+rowplotcount, np.arange(r.shape[0])+0.5,s=scatterpointsize) rowplotcount+=1 if len(col_scatter)>0: col_cluster=True if "col_cluster" in kwargs: col_cluster=kwargs["col_cluster"] for i, lname in enumerate(col_scatter): r=np.array(dfcol[lname]) col_labels.append(np.amin(r)) col_labels.append(np.amax(r)) col_ticks.append(colplotcount) col_ticks.append(colplotcount+tickpos) r=r-np.amin(r) r=r/np.amax(r) r=0.9*r if col_cluster==True: g.ax_col_colors.bar(np.arange(r.shape[0])+0.5,r[leaves_list(g.dendrogram_col.linkage)]+colplotcount,s=scatterpointsize) else: g.ax_col_colors.bar(np.arange(r.shape[0])+0.5,r+colplotcount,s=scatterpointsize) colplotcount+=1 if len(row_bar)>0: row_cluster=True if "row_cluster" in kwargs: row_cluster=kwargs["row_cluster"] for i, lname in enumerate(row_bar): r=np.array(df[lname]) row_labels.append(np.amin(r)) row_labels.append(np.amax(r)) row_ticks.append(rowplotcount) row_ticks.append(rowplotcount+tickpos) r=r-np.amin(r) r=r/np.amax(r) r=0.9*r if row_cluster==True: g.ax_row_colors.barh(y=np.arange(r.shape[0])+0.5, width=r[leaves_list(g.dendrogram_row.linkage)],left=[rowplotcount]*r.shape[0]) else: g.ax_row_colors.barh(r, np.arange(r.shape[0])+0.5,left=[rowplotcount]*r.shape[0]) rowplotcount+=1 if len(col_bar)>0: col_cluster=True if "col_cluster" in kwargs: col_cluster=kwargs["col_cluster"] for i, lname in enumerate(col_bar): r=np.array(dfcol[lname]) col_labels.append(np.amin(r)) col_labels.append(np.amax(r)) col_ticks.append(colplotcount) col_ticks.append(colplotcount+tickpos) if col_cluster==True: g.ax_col_colors.scatter(np.arange(r.shape[0])+0.5,r[leaves_list(g.dendrogram_col.linkage)]/(np.amax(r)*1.1)+colplotcount) else: g.ax_col_colors.scatter(np.arange(r.shape[0])+0.5,r/(np.amax(r)*1.1)+colplotcount) colplotcount+=1 if g.ax_row_colors!=None: g.ax_row_colors.set_xticks(np.arange(len(_row_colors_title))+0.5) g.ax_row_colors.set_xticklabels(_row_colors_title, rotation=90) if g.ax_col_colors!=None: colax_otherside = g.ax_col_colors.twinx() colax_otherside.set_yticks(0.5*(np.arange(len(_col_colors_title))+0.5),labels=_col_colors_title) colax_otherside.grid(False) if g.ax_col_dendrogram!=None: col = g.ax_col_dendrogram.get_position() g.ax_col_dendrogram.set_position([col.x0, col.y0, col.width*0.5, col.height*0.5]) if show_plot_labels==True: rowax_otherside = g.ax_row_colors.twiny() rowax_otherside.invert_xaxis() rowax_otherside.grid(False) rowax_otherside.set_xticks(row_ticks, labels=np.round(row_labels,2), rotation=90, fontsize=8) col_ticks=np.array(col_ticks) g.ax_col_colors.set_yticks(col_ticks, labels=np.round(col_labels,2), fontsize=8) legend_num=0 for _title, colorlut in _row_color_legend.items(): legendhandles=[] for label, color in colorlut.items(): legendhandles.append(Line2D([0], [0], color=color,linewidth=5, label=label)) legend1=g.ax_heatmap.legend(handles=legendhandles, loc="upper left", title=_title, bbox_to_anchor=(1.15, 1-0.2*legend_num)) g.ax_heatmap.add_artist(legend1) legend_num+=1 for _title, colorlut in _col_color_legend.items(): legendhandles=[] for label, color in colorlut.items(): legendhandles.append(Line2D([0], [0], color=color,linewidth=5, label=label)) legend1=g.ax_heatmap.legend(handles=legendhandles, loc="upper left", title=_title, bbox_to_anchor=(1.15, 1-0.2*legend_num)) g.ax_heatmap.add_artist(legend1) legend_num+=1 else: g=sns.clustermap(df,method=method, cbar_kws=dict(orientation='horizontal'), xticklabels=xticklabels, cmap=heatmap_palette, yticklabels=yticklabels, dendrogram_ratio=0.1, figsize=figsize, **kwargs) if color_var>0: cmap = cm.tab20c(np.linspace(0, 1, color_var)) else: cmap = cm.tab20c(np.linspace(0, 1, approx_clusternum+5)) hierarchy.set_link_color_palette([mpl.colors.rgb2hex(rgb[:3]) for rgb in cmap]) """coloring the row dendrogram based on branch numbers crossed with the threshold""" if g.dendrogram_row != None: t=_dendrogram_threshold(g.dendrogram_row.dendrogram,approx_clusternum) den=hierarchy.dendrogram(g.dendrogram_row.linkage, labels = g.data.index, color_threshold=t,ax=g.ax_row_dendrogram, orientation="left") g.ax_row_dendrogram.invert_yaxis() clusters = _get_cluster_classes(den) keys=list(clusters.keys()) ckeys={} i=1 for k in keys: if k=="C0": ckeys[k]="C0" else: ckeys[k]="C"+str(i) i+=1 for c, v in clusters.items(): _c=ckeys[c] for _v in v: cdata["Cluster"].append(_c) cdata["Index"].append(_v) cdata["RGB"].append(matplotlib.colors.to_rgb(c)) """Setting the row dendrogram ends here""" """coloring the col dendrogram based on branch numbers crossed with the threshold""" col_cdata={"Cluster":[],"Index":[],"RGB":[]} if g.dendrogram_col != None: t=_dendrogram_threshold(g.dendrogram_col.dendrogram,approx_clusternum_col) den=hierarchy.dendrogram(g.dendrogram_col.linkage, labels = g.data.columns, color_threshold=t,ax=g.ax_col_dendrogram, orientation="top") #g.ax_col_dendrogram.invert_yaxis() col_clusters = _get_cluster_classes(den) col_cdata={"Cluster":[],"Index":[],"RGB":[]} col_keys=list(col_clusters.keys()) col_ckeys={} i=1 for k in col_keys: if k=="C0": col_ckeys[k]="C0" else: col_ckeys[k]="C"+str(i) i+=1 for c, v in col_clusters.items(): _c=col_ckeys[c] for _v in v: col_cdata["Cluster"].append(_c) col_cdata["Index"].append(_v) col_cdata["RGB"].append(matplotlib.colors.to_rgb(c)) """Setting the col dendrogram ends here""" g.fig.subplots_adjust(bottom=0.175, right=0.70) x0, _y0, _w, _h = g.cbar_pos g.ax_cbar.set_title(ctitle) #g.ax_cbar.tick_params(axis='x', length=5) g.ax_cbar.set_position(pos=[0.80, 0.1, 0.1, 0.02], which="both") if title !="": g.fig.suptitle(title, va="bottom") plt.setp(g.ax_heatmap.xaxis.get_majorticklabels(), rotation=90) _save(save, "complex_clustermap") if show: plt.show() return {"data":g.data2d,"row_clusters":pd.DataFrame(cdata),"col_clusters":pd.DataFrame(col_cdata), "grid":g} def correlation(df: pandas.core.frame.DataFrame, category: Union[str, list] = [], variables: List[~T] = [], method='pearson', palette: str = 'coolwarm', figsize=[6, 6], show_values=False, clustermap_param: dict = {}, ztransform: bool = True, xticklabels=False, yticklabels=False, title: str = '') ‑> Dict[~KT, ~VT]-
Drawing a heatmap with correlations or distances between observations
Parameters
df:pandas DataFramevariables:List, optional- the names of values to calculate correlations
category:strorlist, optional- the names of categorical values to display as color labels
mthod:str- method for correlation/distance calculation. Defalt: "pearson"
palette:str- A colormap name
show_values:bool, optional- Wheter to exhibit the values of fractions/counts/percentages.
clustermap_param:dict, optional- Whether or not to show the figure.
figsize:List[int], optional- The figure size, e.g., [4, 6].
ztransform:bool, optional- Whether to transform values to z-score
xticklabels,yticklabels:bool- Whether to show the label names in the heatmap
Returns
dict
Raises
Notes
References
See Also
Examples
Expand source code
def correlation(df: pd.DataFrame, category: Union[str, list]=[], variables: List=[], method="pearson", palette: str="coolwarm", figsize=[6,6], show_values=False, clustermap_param:dict={}, ztransform: bool=True, xticklabels =False, yticklabels=False, title: str="",)->Dict: """ Drawing a heatmap with correlations or distances between observations Parameters ---------- df : pandas DataFrame variables: List, optional the names of values to calculate correlations category: str or list, optional the names of categorical values to display as color labels mthod: str method for correlation/distance calculation. Defalt: "pearson" palette : str A colormap name show_values: bool, optional Wheter to exhibit the values of fractions/counts/percentages. clustermap_param : dict, optional Whether or not to show the figure. figsize : List[int], optional The figure size, e.g., [4, 6]. ztransform : bool, optional Whether to transform values to z-score xticklabels, yticklabels : bool Whether to show the label names in the heatmap Returns ------- dict Raises ------ Notes ----- References ---------- See Also -------- Examples -------- """ original_index=df.index X, category=_separate_data(df, variables=variables, category=category) # if len(category) !=0: # # if type(category)==str: # category=[category] # #df=df.drop(category, axis=1) # valnames=list(set(df.columns) -set(category)) # X = df[valnames].values # assert X.dtype==float, f"data must contain only float values except {category} column." # # else: # X = df.values # assert X.dtype==float, "data must contain only float values." if ztransform==True: X=zscore(X, axis=0) if method=="pearson": # dmat=Parallel(n_jobs=-1)(delayed(_calc_pearson)(ind, X) for ind in list(it.combinations(range(X.shape[0]), 2))) # dmat=np.array(dmat) # dmat=squareform(dmat) # print(dmat) # dmat+=np.identity(dmat.shape[0]) dmat=np.corrcoef(X) else: dmat=squareform(pdist(X, method)) if method=="pearson": ctitle="Pearson correlation" else: ctitle=method+" distance" if len(category) >0: dfm=pd.DataFrame(data=dmat) colnames=dfm.columns for cat in category: dfm[cat]=df[cat].values res=complex_clustermap(dfm, heatmap_col=colnames, row_colors=category, ztranform=False, xticklabels=xticklabels, yticklabels=yticklabels, figsize=figsize, ctitle=ctitle ) else: res=complex_clustermap(data=dmat, xticklabels=xticklabels, yticklabels=yticklabels, method="ward", cmap=palette, col_cluster=True, row_cluster=True, figsize=figsize, #rasterized=True, #cbar_kws={"label":"Pearson correlation"}, #annot=show_values, **clustermap_param) return res # g.cax.set_ylabel(ctitle, rotation=-90,va="bottom") # plt.setp(g.ax_heatmap.get_yticklabels(), rotation=0) # For y axis # plt.setp(g.ax_heatmap.get_xticklabels(), rotation=90) # For x axis # return {"grid":g} def dotplot(df: pandas.core.frame.DataFrame, y: str = '', x: str = '', dfc=Empty DataFrame Columns: [] Index: [], scaling: float = 10, color_val: str = '', size_val: str = '', highlight: str = '', color_title: str = '', size_title: str = '', figsize: list = [], save: str = '', threshold: float = 1.3010299956639813, row_clustering: bool = True, xtickrotation: float = 90, column_order: list = [], colorpalette='coolwarm', show: bool = False, title: str = '', row: str = '', col: str = '') ‑> Dict[str, matplotlib.axes._axes.Axes]-
Drawing a dotplot that can represent two different variables as dot sizes and colors on a regular grid. This function is assumed to plot GO enrichment analysis with multiple gene sets.
Parameters
df:pandas DataFrame- dataframe containing two categories and corresponding values (such as p values and odds ratio). e.g.: Cluster Condensate pval odds FDR 54 C1 Cajal body -0.000000 0.000000 -0.000000 55 C1 DNA repair focus -0.000000 0.000000 -0.000000 56 C1 DNA replication condensate -0.000000 0.000000 -0.000000 57 C1 P-body -0.000000 0.000000 -0.000000 58 C1 PML body -0.000000 0.000000 -0.000000
row:string- the column name of a category that is going to be placed in the row of the dotplot
col:string- the column name of a category that is going to be placed in the column of the dotplot
color_val:string- The column name for the values represented as dot colors.
size_val:string- The column name for the values represented as dot sizes.
scaling:float- The scale of dots. If resulting dots are too large (or small), you can reduce (or increase) dot sizes by adjusting this value.
highlight:string- A dictionary to set color labels to leaves. The key is the name of the color label. The value is the list of RGB color codes, each corresponds to the color of a leaf. e.g., {"color1":[[1,0,0,1], ....]}
size_title:string- The title for size values. If not set, "size_val" will be used.
color_title:string- The title for color values. If not set, "color_val" will be used.
show:bool- Whether or not to show the figure.
Returns
axes:dict {"axes1":ax1,"axes2":ax2,"axes3":ax3}
Raises
Notes
References
See Also
Examples
Expand source code
def dotplot(df: pd.DataFrame, y: str="", x: str="", dfc=pd.DataFrame(), scaling: float=10, color_val: str="", size_val: str="", highlight: str="", color_title: str="", size_title: str="", figsize: list=[], save: str="", threshold: float=-np.log10(0.05), row_clustering: bool=True, xtickrotation: float=90, column_order: list=[], colorpalette="coolwarm", show: bool=False, title: str="", row: str="", col: str="", ) -> Dict[str, plt.Axes]: """ Drawing a dotplot that can represent two different variables as dot sizes and colors on a regular grid. This function is assumed to plot GO enrichment analysis with multiple gene sets. Parameters ---------- df : pandas DataFrame dataframe containing two categories and corresponding values (such as p values and odds ratio). e.g.: Cluster Condensate pval odds FDR 54 C1 Cajal body -0.000000 0.000000 -0.000000 55 C1 *DNA repair focus -0.000000 0.000000 -0.000000 56 C1 *DNA replication condensate -0.000000 0.000000 -0.000000 57 C1 P-body -0.000000 0.000000 -0.000000 58 C1 PML body -0.000000 0.000000 -0.000000 row: string the column name of a category that is going to be placed in the row of the dotplot col: string the column name of a category that is going to be placed in the column of the dotplot color_val : string The column name for the values represented as dot colors. size_val : string The column name for the values represented as dot sizes. scaling: float The scale of dots. If resulting dots are too large (or small), you can reduce (or increase) dot sizes by adjusting this value. highlight : string A dictionary to set color labels to leaves. The key is the name of the color label. The value is the list of RGB color codes, each corresponds to the color of a leaf. e.g., {"color1":[[1,0,0,1], ....]} size_title : string The title for size values. If not set, "size_val" will be used. color_title : string The title for color values. If not set, "color_val" will be used. show : bool Whether or not to show the figure. Returns ------- axes: dict {"axes1":ax1,"axes2":ax2,"axes3":ax3} Raises ------ Notes ----- References ---------- See Also -------- Examples -------- """ if x !="" and y!="": row=y col=x if size_val!="": _df=df.pivot_table(index=col,columns=row,values=size_val) if len(column_order)>0: _df=_df.loc[column_order] else: _df=_df.sort_index(axis=0,key=natsort_keygen()) _df=_df.fillna(0) if color_val!="": dfc=df.pivot_table(index=col,columns=row,values=color_val) if len(column_order)>0: dfc=dfc.loc[column_order] else: dfc=dfc.sort_index(axis=0,key=natsort_keygen()) dfc=dfc.fillna(0) if highlight !="": dfh=df.pivot_table(index=col,columns=row,values=highlight) if len(column_order)>0: dfh=dfh.loc[column_order] else: dfh=dfh.sort_index(axis=0,key=natsort_keygen()) dfh=dfh.fillna(0) if row_clustering==True: Y = fcl.linkage(_df.to_numpy().T, method='ward', metric='euclidean') Z1 = sch.dendrogram(Y,no_plot=True) idx1 = Z1['leaves'] reorder=_df.columns[idx1] _df=_df[reorder] if color_val!="": dfc=dfc[reorder] if highlight !="": dfh=dfh[reorder] _x=_df.index _y=_df.columns mat=_df.to_numpy() minsize=np.amin(mat) maxsize=np.amax(mat) else: _x=df.index _y=df.columns mat=df.to_numpy() minsize=np.amin(mat) maxsize=np.amax(mat) #if minsize==0: #mat=mat+maxsize*0.01 #minsize=np.amin(mat) maxsize=np.round(maxsize) middle0=np.round((minsize+maxsize)/3) middle1=np.round(2*(minsize+maxsize)/3) #plt.rcParams["figure.figsize"] = [7.50, 3.50] #plt.rcParams["figure.autolayout"] = True #x = np.arange(len(_x)) #y = np.arange(len(_y)) #X, Y = np.meshgrid(x, y) xy=[ [i,j] for i in range(len(_x)) for j in range(len(_y))] #num = 1000 sizes = [mat[i,j]*scaling for i in range(len(_x)) for j in range(len(_y))] edge_colors=[] if highlight !="": hmat=dfh.to_numpy() hvals = [hmat[i,j]*scaling for i in range(len(_x)) for j in range(len(_y))] for s in hvals: if s>=threshold*scaling: edge_colors.append("magenta") else: edge_colors.append("gray") else: for s in sizes: if s>=threshold*scaling: edge_colors.append("magenta") else: edge_colors.append("gray") if len(dfc) !=0: viridis = cm.get_cmap(colorpalette, 12) cmat=dfc.to_numpy() cmat[cmat==np.inf]=0 _cmat=cmat/np.amax(cmat) _colors = [viridis(_cmat[i,j]) for i in range(len(_x)) for j in range(len(_y))] else: _colors = [[0,1,0,1] for i in range(len(_x)) for j in range(len(_y))] #print(sizes) #xy = 10 * np.random.random((num, 2)) #xy=XY #patches = [plt.Circle(center, size) for center, size in zip(xy, sizes)] #fig, ax = plt.subplots(ncols=2, gridspec_kw={'width_ratios': [8, 2]}) if len(figsize)==0: figsize=[mat.shape[0]*0.5+2,mat.shape[1]*0.5+1] fig = plt.figure(figsize=figsize) #fig.set_figheight(6) #fig.set_figwidth(6) ax1 = plt.subplot2grid(shape=(10, 6), loc=(0, 0), colspan=4, rowspan=10) ax2 = plt.subplot2grid(shape=(10, 6), loc=(1, 4), colspan=2, rowspan=4) ax3 = plt.subplot2grid(shape=(10, 6), loc=(6, 4), colspan=2, rowspan=1) collection = mc.CircleCollection(sizes, edgecolors=edge_colors, offsets=xy, transOffset=ax1.transData, facecolors=_colors, linewidths=2) ax1.add_collection(collection) ax1.margins(0.1) ax1.set_xlim(-0.5,len(_x)-0.5) ax1.set_xticks(np.arange(len(_x))) ax1.set_xticklabels(_x,rotation=xtickrotation) ax1.set_yticks(np.arange(len(_y))) ax1.set_yticklabels(_y, rotation=0) if color_title=="": color_title=color_val if len(dfc) !=0: norm = mpl.colors.Normalize(vmin=np.min(cmat), vmax=np.amax(cmat)) cb1 = mpl.colorbar.ColorbarBase(ax3, cmap=viridis, norm=norm, orientation='horizontal') cb1.set_label(color_title) #ax[1]=fig.add_axes([1,0.3,0.1,1]) lxy=[[0.5, i*0.5] for i in range(3)] collection2 = mc.CircleCollection([middle0*scaling,middle1*scaling, maxsize*scaling], offsets=lxy, transOffset=ax2.transData, facecolors='lightgray', edgecolors="gray") ax2.add_collection(collection2) ax2.axis('off') ax2.margins(0.3) for text, (x, y) in zip([middle0,middle1, maxsize], lxy): ax2.text(x+0.01, y,str(text), ha="left",va="center",color="black" ) if size_title=="": size_title=size_val ax2.text(0.5,-0.5, size_title,va="center",ha="center") #ax[1].set_yticks(np.arange(3)) #ax[1].set_yticklabels([minsize,middle, maxsize], rotation=0) #plt.tight_layout() plt.subplots_adjust(left=0.3,bottom=0.2) #plt.tight_layout() if title !="": fig.suptitle(title) _save(save, "dotplot") if show==True: plt.show() return {"axes1":ax1,"axes2":ax2,"axes3":ax3} def radialtree(df: pandas.core.frame.DataFrame, n_clusters: int = 3, x: str = '', variables: List[~T] = [], category: Union[str, List[str]] = [], ztransform: bool = True, save: str = '', distance_method='euclidean', tree_method='ward', title: str = '', y: list = [], linewidth: float = 1, figsize: Optional[list] = None, **kwargs) ‑> Dict[~KT, ~VT]-
Drawing a radial dendrogram with color labels.
Parameters
df:pandas DataFrame- A wide format data.
n_clusters:int, optional(default: 3)- Approximate number of clusters to produce
x:str, optional- the name of columns containing sample names. If not provided, the index will be considered sample names.
variables:list, optional- the name of columns containing variables to calculate the distances between samples
category:strorlistofstr- the column name of a category that is going to presented as colors around the dendrogram.
ztransform:bool=True,save:str="",
distance_method="euclidean", tree_method="ward",
show:bool- Whether or not to show the figure.
fontsize:float- A float to specify the font size
figsize:[x, y] array-like- 1D array-like of floats to specify the figure size
palette:string- Matlab colormap name.
Returns
dict:{"axes":ax, "clusters": clusters}
Raises
Notes
References
See Also
Examples
Expand source code
def radialtree(df: pd.DataFrame, n_clusters: int=3, x: str="", variables: List=[], category: Union[str, List[str]]=[], ztransform: bool=True, save: str="", distance_method="euclidean", tree_method="ward", title: str="", y: list=[],linewidth: float=1,figsize: Optional[list]=None, **kwargs) -> Dict: """ Drawing a radial dendrogram with color labels. Parameters ---------- df : pandas DataFrame A wide format data. n_clusters: int, optional (default: 3) Approximate number of clusters to produce x: str, optional the name of columns containing sample names. If not provided, the index will be considered sample names. variables: list, optional the name of columns containing variables to calculate the distances between samples category: str or list of str the column name of a category that is going to presented as colors around the dendrogram. ztransform: bool=True, save: str="", distance_method="euclidean", tree_method="ward", show : bool Whether or not to show the figure. fontsize : float A float to specify the font size figsize : [x, y] array-like 1D array-like of floats to specify the figure size palette : string Matlab colormap name. Returns ------- dict: {"axes":ax, "clusters": clusters} Raises ------ Notes ----- References ---------- See Also -------- Examples -------- """ if len(y)!=0: variables=y if x !="": _labels=df[x] df=df.drop(x, axis=1) else: _labels=df.index X, category=_separate_data(df, variables=variables, category=category) category_df=df[category] # if len(variables)!=0 and len(category)!=0: # if type(category)==str: # category=[category] # category_df=df[category] # df=df[variables] # X = df.values # #print(X) # assert X.dtype==float, f"{x} columns must contain only float values." # # # elif len(category) !=0: # if type(category)==str: # category=[category] # category_df=df[category] # df=df.drop(category, axis=1) # X = df.values # #print(X) # assert X.dtype==float, f"data must contain only float values except {category} column." # # else: # X = df.values # assert X.dtype==float, "data must contain only float values." if ztransform==True: X=zscore(X, axis=0) D=squareform(pdist(X,metric=distance_method)) Y = sch.linkage(D, method=tree_method) Z = sch.dendrogram(Y,labels=_labels,no_plot=True) t=_dendrogram_threshold(Z, n_clusters) _fig, _ax =plt.subplots() Z=sch.dendrogram(Y, labels = _labels, color_threshold=t, ax=_ax) #xticks=set(_ax.get_xticks()) plt.close(_fig) sample_classes={k: list(category_df[k]) for k in category_df.columns} ax=_radialtree2(Z, sample_classes=sample_classes, addlabels=False, linewidth=linewidth, figsize=figsize, **kwargs) if title !="": ax.set_title(title) _save(save, "radialtree") clusters = _get_cluster_classes(Z) return {"axes":ax, "clusters":clusters} def triangle_heatmap(df, grid_pos: list = [], grid_labels: list = [], show: bool = False, save: str = '', title: str = '') ‑> dict-
Creating a heatmap with 45 degree rotation.
Parameters
df:pandas DataFramegrid_pos:list- the column name of a category that is going to be placed in the row of the dotplot
grid_labels:list- the column name of a category that is going to be placed in the column of the dotplot
show:bool- Whether or not to show the figure.
Returns
dict {"axes": ax}
Raises
Notes
References
See Also
Examples
Expand source code
def triangle_heatmap(df, grid_pos: list=[], grid_labels: list=[], show: bool=False, save: str="",title: str="")-> dict: """ Creating a heatmap with 45 degree rotation. Parameters ---------- df : pandas DataFrame grid_pos: list the column name of a category that is going to be placed in the row of the dotplot grid_labels: list the column name of a category that is going to be placed in the column of the dotplot show : bool Whether or not to show the figure. Returns ------- dict {"axes": ax} Raises ------ Notes ----- References ---------- See Also -------- Examples -------- """ genes=df.index fig, ax = plt.subplots(figsize=[8,6]) dmat=df.to_numpy() D = dmat N = dmat.shape[0] a=np.tril(np.zeros([N,N])-1000000, k=-1) # Get the lower triangle of the matrix. C = np.triu(D)+a # Mask the upper triangle. C = np.ma.masked_array(C, C == -1000000) # Set the diagonal to zero. for i in range(N): C[i, i] = 0 # Transformation matrix for rotating the heatmap. A = np.array([(y, x) for x in range(N, -1, -1) for y in range(N + 1)]) t = np.array([[2**(-0.5), 2**(-0.5)], [-2**(-0.5), 2**(-0.5)]]) A = np.dot(A, t) #t_ = np.array([[2**(-0.5), -2**(-0.5)], [2**(-0.5), 2**(-0.5)]]) # -1.0 correlation is blue, 0.0 is white, 1.0 is red. cmap = plt.cm.Reds #norm = mp.colors.BoundaryNorm(np.linspace(0, 10, 14), cmap.N) # This MUST be before the call to pl.pcolormesh() to align properly. ax.set_xticks([]) ax.set_yticks([]) X = A[:, 1].reshape(N + 1, N + 1) Y = A[:, 0].reshape(N + 1, N + 1) caxes = plt.pcolormesh(X, Y, np.flipud(C), axes=ax, cmap=cmap, rasterized=True) x1s=[0] if len(grid_pos)>0: for i, grid in enumerate(grid_pos): x0=(grid)/(2**(0.5)) x1=(grid)*(2**0.5) y0=(grid)/(2**(0.5)) y1=0 ax.plot([x0, x1], [y0, y1],color='gray', linewidth=1) ax.plot([(grid)*(2**0.5), (grid)*(2**0.5)+(N-grid)/(2**(0.5))], [0, (N-grid)/(2**(0.5))], color='gray', linewidth=1) x1s.append(x1) x1s.append(N*2**0.5) if len(genes) >0: leng=0 for i, g in enumerate(genes): #ax.plot([(i+0.5)*(2**0.5), (i+0.5)*(2**0.5)], [-2**0.5, -(N/10)*2**0.5/2],color='b', linewidth=0.5) ax.text((i+0.5)*(2**0.5), -(N/10)*2**0.5/2*1.01, g, rotation=90,ha='center', va='top', fontsize="small") leng+=len(g) leng=leng/len(genes) else: leng=0 spacing=(N/10)*leng*0.4 rect=Rectangle([0, -spacing],N*(2**0.5), spacing, color='whitesmoke' , alpha=1, linewidth=0) ax.add_patch(rect) rect=Rectangle([0, -(N/10)*2**0.5/2],N*(2**0.5), (N/10)*2**0.5/2, color='dimgray' , alpha=1, linewidth=0) ax.add_patch(rect) if len(grid_labels)>0: for i in range(len(x1s)-1): if i%2==1: rect=Rectangle([x1s[i], -spacing],np.abs(x1s[i] - x1s[i+1]), spacing, color='silver' , alpha=0.3, linewidth=0) ax.add_patch(rect) else: rect=Rectangle([x1s[i], -spacing],np.abs(x1s[i] - x1s[i+1]), spacing, color='lavender' , alpha=0.3, linewidth=0) ax.add_patch(rect) x=(x1s[i]+x1s[i+1])/2 ax.text(x,-(N/10)*2**0.5/4, grid_labels[i], rotation=90,ha='center', va='center', color="w") cb = plt.colorbar(caxes, ax=ax, shrink=0.75) ax.spines['top'].set_visible(False) ax.spines['right'].set_visible(False) ax.spines['bottom'].set_visible(False) ax.spines['left'].set_visible(False) if save!="": if save.endswith(".pdf") or save.endswith(".png") or save.endswith(".svg"): plt.savefig(save) else: plt.savefig(save+"_complexheatmap.pdf") if show: plt.show() return {"axes": ax} def violinplot(df: pandas.core.frame.DataFrame, x: str, y: str, pairs: list = [], test: str = 'ttest_ind', alternative: str = 'two-sided', significance: str = 'numeric', significance_ranges: Dict[str, float] = {'*': 1.3010299956639813, '**': 4, '***': 10}, swarm: bool = False, xorder: list = [], equal_var: bool = False, yunit: str = '', title: str = '', save: str = '', ax: Optional[matplotlib.axes._axes.Axes] = None, **kwargs)-
Draw a boxplot with a statistical test
Parameters
df:pandas DataFrame- x,y: str
- names of variables in data
pairs:list, optional- Category pairs for the statistical test. Examples: [["Adelie","Chinstrap" ], ["Gentoo","Chinstrap" ], ["Adelie","Gentoo" ]]
test:str, optional- Method name for the statistical test. Defalt: ttest_ind Available methods: ["ttest_ind", "ttest_rel", "kruskal", "mannwhitneyu", "wilcoxon", "brunnermunzel", "median_test"]
alternative:str ['two-sided', 'less', 'greater'], optional- Defines the alternative hypothesis. Defalt: "two-sided"
show:bool, optional- Whether or not to show the figure.
significance:str ['numeric', 'symbol'], optional- How to show the significance. 'numeric' will show -log10(p values) in the plot and 'symbol' will represent significance as asterisks.
significance_ranges:dict, optional- thresholds of -log10(p values) that each asterisk number represents. Ignored when significance="numeric". example: {"":-np.log10(0.05),"":4,"**":10}
swarm:bool, optional- Whether or not to superpose a swarm plot. Not recommended if the sample size is too large.
xorder:list, optional- The order of x axis labels
equal_var:bool, optional- Related to ttest_ind method. The default is True, which will produce a p value equal to t-test in R.
kwargs:any options accepted by scipy statistical test functions
Returns
dict("p values":pvalues,"axes":ax)
Raises
Notes
References
See Also
Examples
Expand source code
def violinplot(df: pd.DataFrame, x: str, y: str, pairs: list=[], test: str="ttest_ind", alternative: str="two-sided", significance: str="numeric", significance_ranges: Dict[str, float]={"*":-np.log10(0.05),"**":4,"***":10}, swarm: bool=False, xorder: list=[], equal_var: bool=False, yunit: str="", title: str="", save: str="", ax: Optional[plt.Axes]=None,**kwargs): """ Draw a boxplot with a statistical test Parameters ---------- df : pandas DataFrame x,y: str names of variables in data pairs: list, optional Category pairs for the statistical test. Examples: [["Adelie","Chinstrap" ], ["Gentoo","Chinstrap" ], ["Adelie","Gentoo" ]] test: str, optional Method name for the statistical test. Defalt: ttest_ind Available methods: ["ttest_ind", "ttest_rel", "kruskal", "mannwhitneyu", "wilcoxon", "brunnermunzel", "median_test"] alternative: str ['two-sided', 'less', 'greater'], optional Defines the alternative hypothesis. Defalt: "two-sided" show : bool, optional Whether or not to show the figure. significance: str ['numeric', 'symbol'], optional How to show the significance. 'numeric' will show -log10(p values) in the plot and 'symbol' will represent significance as asterisks. significance_ranges: dict, optional thresholds of -log10(p values) that each asterisk number represents. Ignored when significance="numeric". example: {"*":-np.log10(0.05),"**":4,"***":10} swarm: bool, optional Whether or not to superpose a swarm plot. Not recommended if the sample size is too large. xorder: list, optional The order of x axis labels equal_var: bool, optional Related to ttest_ind method. The default is True, which will produce a p value equal to t-test in R. kwargs: any options accepted by scipy statistical test functions Returns ------- dict("p values":pvalues,"axes":ax) Raises ------ Notes ----- References ---------- See Also -------- Examples -------- """ tests=["ttest_ind","ttest_rel","kruskal","mannwhitneyu","wilcoxon","brunnermunzel","median_test"] if not test in tests: raise Exception("Available tests are "+", ".join(tests)) import scipy.stats as stats if len(xorder)==0: xorder=sorted(list(set(df[x]))) pvals=[] for p1,p2 in pairs: statstest=getattr(stats, test) if test=="wilcoxon" or test=="ttest_rel": _, pval,_=statstest(df[y][df[x]==p1],df[y][df[x]==p2],alternative=alternative,**kwargs) elif test=="median_test": _, pval,_,_=statstest(df[y][df[x]==p1],df[y][df[x]==p2],alternative=alternative,**kwargs) elif test=="ttest_ind": _, pval=statstest(df[y][df[x]==p1],df[y][df[x]==p2],alternative=alternative,equal_var=equal_var,**kwargs) else: _, pval=statstest(df[y][df[x]==p1],df[y][df[x]==p2],alternative=alternative,**kwargs) p1ind=xorder.index(p1) p2ind=xorder.index(p2) if pval==0: pval=np.inf else: pval=-np.log10(pval) pvals.append([np.abs(p2ind-p1ind), np.amin([p2ind, p1ind]),np.amax([p2ind, p1ind]), pval]) pvals = sorted(pvals, key = lambda x: (x[0], x[1])) fig, ax=plt.subplots() sns.violinplot(data=df, x=x,y=y,inner="quartile") if swarm==True: sns.swarmplot(data=df, x=x,y=y,color="black",alpha=0.5) ymax=np.amax(df[y]) newpvals={} for i, pval in enumerate(pvals): plt.plot([pval[1],pval[2]], [ymax*(1.05+i*0.05),ymax*(1.05+i*0.05)], color="black") p=np.round(pval[-1],2) newpvals[xorder[pval[1]]+"_"+xorder[pval[2]]]=p if significance=="numeric": annotate="-log10(p)="+str(p) elif significance=="symbol": keys=sorted(significance_ranges.keys()) annotate="NA" for j in range(len(keys)): if j==0: if p <= significance_ranges[keys[j]]: annotate="" break else: if significance_ranges[keys[j-1]] < p <=significance_ranges[keys[j]]: annotate=keys[i] break if annotate=="NA": annotate=keys[-1] plt.text((pval[1]+pval[2])/2, ymax*(1.055+i*0.05), annotate) if significance=="symbol": ax.annotate("\n".join(["{}: p < {:.2E}".format(k, 10**(-significance_ranges[k])) for k in keys]), xy=(0.9,0.9), xycoords='axes fraction', textcoords='offset points', size=12, bbox=dict(boxstyle="round", fc=(0.9, 0.9, 0.9), ec="none")) plt.subplots_adjust(right=0.850) if yunit!="": ax.text(0, 1, "({})".format(yunit), transform=ax.transAxes, ha="right") _save(save, "violin") return {"p values":newpvals,"axes":ax}