[2]:

import os,sys
%matplotlib inline
import matplotlib.pylab as plt
import pickle
import numpy as np
plt.rcParams['figure.dpi'] = 100
plt.rcParams['savefig.dpi']=300
sys.path.append(os.path.expanduser("~/Projects/Github/PyComplexHeatmap"))
from PyComplexHeatmap import *
use_pch_style() # or plt.style.use('default') to restore default style
# plt.rcParams
# import matplotlib; print(matplotlib.__version__)

[3]:

#set font to Arial using the following code
plt.rcParams['font.family']='sans serif'
plt.rcParams['font.sans-serif']='Arial'
# set pdf.fonttype to 42
plt.rcParams['pdf.fonttype']=42

Generate dataset¶

[3]:

#Generate example dataset (random)
df = pd.DataFrame(['GroupA'] * 5 + ['GroupB'] * 5, columns=['AB'])
df['CD'] = ['C'] * 3 + ['D'] * 3 + ['G'] * 4
df['EF'] = ['E'] * 6 + ['F'] * 2 + ['H'] * 2
df['F'] = np.random.normal(0, 1, 10)
df.index = ['sample' + str(i) for i in range(1, df.shape[0] + 1)]
df_box = pd.DataFrame(np.random.randn(10, 4), columns=['Gene' + str(i) for i in range(1, 5)])
df_box.index = ['sample' + str(i) for i in range(1, df_box.shape[0] + 1)]
df_bar = pd.DataFrame(np.random.uniform(0, 10, (10, 2)), columns=['TMB1', 'TMB2'])
df_bar.index = ['sample' + str(i) for i in range(1, df_box.shape[0] + 1)]
df_scatter = pd.DataFrame(np.random.uniform(0, 10, 10), columns=['Scatter'])
df_scatter.index = ['sample' + str(i) for i in range(1, df_box.shape[0] + 1)]
df_heatmap = pd.DataFrame(np.random.randn(30, 10), columns=['sample' + str(i) for i in range(1, 11)])
df_heatmap.index = ["Fea" + str(i) for i in range(1, df_heatmap.shape[0] + 1)]
df_heatmap.iloc[1, 2] = np.nan

[4]:

# add a missing value to sample4
df_heatmap.loc['Fea4','sample4']=np.nan
df_box.loc['sample4','Gene4']=np.nan
df_box

[4]:

	Gene1	Gene2	Gene3	Gene4
sample1	-1.141544	-1.854671	0.700472	-1.453638
sample2	-1.589981	1.318247	-0.392954	0.854834
sample3	2.473380	1.425263	1.838210	-0.299281
sample4	0.615659	-1.128731	-0.918126	NaN
sample5	-0.326837	0.456930	1.591883	-2.251957
sample6	-0.169985	0.294284	-1.784905	0.175566
sample7	0.267317	0.164426	-0.332801	-1.572528
sample8	-1.620649	-0.766258	0.555885	-0.579876
sample9	-0.769411	-2.204664	-0.712077	-1.359158
sample10	0.539084	-0.762180	-1.011151	1.060917

Add selected rows labels¶

[5]:

#Annotate the rows with average > 0.3
df_rows = df_heatmap.apply(lambda x:x.name if x.sample4 > 0.5 else None,axis=1)
df_rows=df_rows.to_frame(name='Selected')
df_rows['XY']=df_rows.index.to_series().apply(lambda x:'A' if int(x.replace('Fea',''))>=15 else 'B')

row_ha = HeatmapAnnotation(
            Scatter=anno_scatterplot(df_heatmap.sample4.apply(lambda x:round(x,2)),
                            height=12,cmap='jet',legend=False,grid=True),
            Line=anno_lineplot(df_heatmap.sample4.apply(lambda x:round(x,2)),
                            height=12,colors='red',linewidth=2,legend=False),
            Bar=anno_barplot(df_heatmap.sample4.apply(lambda x:round(x,2)),
                            height=15,cmap='rainbow',legend=False),
            selected=anno_label(df_rows,colors='red',relpos=(-0.05,0.4)),
            label_kws={'rotation':30,'horizontalalignment':'left','verticalalignment':'bottom'},
            axis=0,verbose=0)

col_ha = HeatmapAnnotation(
            label=anno_label(df.AB, merge=True,rotation=10,
                             arrowprops = dict(visible=False,)
                            ), #visible in arrowprops can control whether to show the arrow
            AB=anno_simple(df.AB,add_text=True),axis=1,
            CD=anno_simple(df.CD,add_text=True),
            EF=anno_simple(df.EF,add_text=True,
                            legend_kws={'frameon':True}),
            G=anno_boxplot(df_box, cmap='jet',legend=False,grid=True),
            verbose=0)

print(np.nanmin(df_heatmap),np.nanmax(df_heatmap))

plt.figure(figsize=(5.5, 6.5))
cm = ClusterMapPlotter(
        data=df_heatmap, top_annotation=col_ha,right_annotation=row_ha,
        col_cluster=True,row_cluster=True,
        col_split=df.AB,row_split=2, z_score=0,vmin=-2.2,vmax=2.3,
        col_split_gap=0.5,row_split_gap=0.8,
        label='values',row_dendrogram=True,col_dendrogram=False,row_dendrogram_size=15,
        show_rownames=False,show_colnames=True,
        tree_kws={'row_cmap': 'Set1'},verbose=0,legend_gap=5,
        cmap='RdYlBu_r',bezier=True,dotsize=2,
        xticklabels_kws=dict(labelrotation=90,labelcolor='blue',labelsize=14,grid_color='red',bottom=True))
# for ax in cm.top_annotation.axes[-1,:]:
#     ax.cla()
plt.savefig("example0.pdf", bbox_inches='tight')
plt.show()
print(cm.kwargs['vmin'],cm.kwargs['vmax'],cm.legend_kws)

-3.195712714460433 3.3567559056050413

../_images/notebooks_advanced_usage_6_1.png

-2.2 2.3 {'vmin': -2.2, 'vmax': 2.3, 'center': None}

[6]:

cm.heatmap_axes

[6]:

array([[<Axes: >, <Axes: >],
       [<Axes: >, <Axes: >]], dtype=object)

[7]:

cm.ax_heatmap.xaxis.get_tick_params()

[7]:

{'rotation': 90,
 'left': True,
 'right': False,
 'labelleft': False,
 'labelright': False,
 'gridOn': False,
 'labelcolor': 'blue',
 'labelsize': 14,
 'grid_color': 'red'}

Add annotations on the top of heatmap cells¶

[8]:

#Annotate the rows with average > 0.3
df_rows = df_heatmap.apply(lambda x:x.name if x.sample4 > 0.5 else None,axis=1)
df_rows=df_rows.to_frame(name='Selected')
df_rows['XY']=df_rows.index.to_series().apply(lambda x:'A' if int(x.replace('Fea',''))>=15 else 'B')

row_ha = HeatmapAnnotation(
            S4=anno_simple(df_heatmap.sample4.apply(lambda x:round(x,2) if not pd.isna(x) else ''),
                       add_text=True,height=10,legend=False,
                       text_kws={'rotation':0,'fontsize':10,'color':'blue'}),
            # Scatter=anno_scatterplot(df_heatmap.sample4.apply(lambda x:round(x,2)),
            #                  height=10),
            Test=anno_barplot(df_heatmap.sample4.apply(lambda x:round(x,2)),
                        height=18,cmap='rainbow',grid=True),
            selected=anno_label(df_rows,colors='red'),
            axis=0,verbose=0,#wgap=4,
            label_kws={'rotation':0,'horizontalalignment':'left',
                                        'verticalalignment':'bottom'})

col_ha = HeatmapAnnotation(
            label=anno_label(df.AB, merge=True,rotation=15),
            AB=anno_simple(df.AB,add_text=True),axis=1,
            CD=anno_simple(df.CD,add_text=True,text_kws=dict(bbox={'boxstyle':'Circle','edgecolor':'white','fill':False},fontsize=8),
                           height=4.5),
            EF=anno_simple(df.EF,add_text=True,
                            legend_kws={'frameon':False}),
            Exp=anno_boxplot(df_box, cmap='turbo',grid=True),
            verbose=0,#hgap=2
            ) #verbose=0 will turn off the log.

print(df.head())
print(df_box.mean(axis=1))
print(df_heatmap.head())
plt.figure(figsize=(6, 8))
cm = ClusterMapPlotter(
            data=df_heatmap, top_annotation=col_ha,right_annotation=row_ha,
            col_split=df.AB,
            row_split=df_rows.XY,
            #col_split_gap=3.5,row_split_gap=2.5,
            col_cluster=True,row_cluster=True,
            label='values',row_dendrogram=False,
            show_rownames=True,show_colnames=True,
            verbose=0,legend_gap=5,#legend_hpad=10,legend_vpad=5,
            annot=False,fmt='.1g',linewidths=0.05,linecolor='gold',cmap='RdYlBu_r',
            xticklabels_kws={'labelrotation':-45,'labelcolor':'blue'},
            yticklabels_kws=dict(labelcolor='red'),#subplot_gap=8
)
#subplot_gap controls the gap between main heatmap and column or row annotations

plt.show()
print(cm.row_order)
print(cm.col_order)

             AB CD EF         F
sample1  GroupA  C  E  0.362690
sample2  GroupA  C  E -0.982457
sample3  GroupA  C  E  1.067007
sample4  GroupA  D  E -1.318330
sample5  GroupA  D  E  0.046514
sample1    -0.937345
sample2     0.047537
sample3     1.359393
sample4    -0.477066
sample5    -0.132495
sample6    -0.371260
sample7    -0.368397
sample8    -0.602725
sample9    -1.261328
sample10   -0.043332
dtype: float64
       sample1   sample2   sample3   sample4   sample5   sample6   sample7  \
Fea1 -0.256096  1.294337 -0.085543 -0.243972  1.303379  0.755107  0.361045
Fea2  1.735609  0.316170       NaN  1.830172  0.544176  1.372135  0.351998
Fea3  0.733865 -0.798262  0.775947  0.002780 -0.826095 -0.162563 -0.860632
Fea4  0.781866 -0.127187  0.703247       NaN  0.164684  0.433726 -0.256959
Fea5  3.356756 -0.846319 -0.207551 -0.409454  0.172251  0.093536 -0.893011

       sample8   sample9  sample10
Fea1  0.330225  1.236336  0.429788
Fea2 -0.153091  0.981548 -0.476542
Fea3 -0.143750  0.184459  0.851115
Fea4 -0.030188 -1.230343  1.477730
Fea5 -1.336817 -0.312624  1.645297

../_images/notebooks_advanced_usage_10_1.png

[['Fea24', 'Fea18', 'Fea30', 'Fea27', 'Fea20', 'Fea17', 'Fea19', 'Fea21', 'Fea28', 'Fea15', 'Fea26', 'Fea23', 'Fea25', 'Fea22', 'Fea16', 'Fea29'], ['Fea6', 'Fea8', 'Fea14', 'Fea2', 'Fea4', 'Fea3', 'Fea5', 'Fea12', 'Fea9', 'Fea10', 'Fea11', 'Fea13', 'Fea1', 'Fea7']]
[['sample2', 'sample3', 'sample4', 'sample1', 'sample5'], ['sample6', 'sample10', 'sample7', 'sample8', 'sample9']]

Only plot the annotations¶

[9]:

df = pd.DataFrame(['AAAA1'] * 5 + ['BBBBB2'] * 5, columns=['AB'])
df['CD'] = ['C'] * 3 + ['D'] * 3 + ['G'] * 4
df['F'] = np.random.normal(0, 1, 10)
df.index = ['sample' + str(i) for i in range(1, df.shape[0] + 1)]
df_box = pd.DataFrame(np.random.randn(10, 4), columns=['Gene' + str(i) for i in range(1, 5)])
df_box.index = ['sample' + str(i) for i in range(1, df_box.shape[0] + 1)]
df_bar = pd.DataFrame(np.random.uniform(0, 10, (10, 2)), columns=['TMB1', 'TMB2'])
df_bar.index = ['sample' + str(i) for i in range(1, df_box.shape[0] + 1)]
df_scatter = pd.DataFrame(np.random.uniform(0, 10, 10), columns=['Scatter'])
df_scatter.index = ['sample' + str(i) for i in range(1, df_box.shape[0] + 1)]
df_bar1 = pd.DataFrame(np.random.uniform(0, 10, (10, 2)), columns=['T1-A', 'T1-B'])
df_bar1.index = ['sample' + str(i) for i in range(1, df_box.shape[0] + 1)]
df_bar2 = pd.DataFrame(np.random.uniform(0, 10, (10, 2)), columns=['T2-A', 'T2-B'])
df_bar2.index = ['sample' + str(i) for i in range(1, df_box.shape[0] + 1)]
df_bar3 = pd.DataFrame(np.random.uniform(0, 10, (10, 2)), columns=['T3-A', 'T3-B'])
df_bar3.index = ['sample' + str(i) for i in range(1, df_box.shape[0] + 1)]
df_bar3.iloc[5,0]=np.nan
df_bar4 = pd.DataFrame(np.random.uniform(0, 10, (10, 1)), columns=['T4'])
df_bar4.index = ['sample' + str(i) for i in range(1, df_box.shape[0] + 1)]
df_bar4.iloc[7,0]=np.nan
print(df)
print(df_box.head())
print(df_scatter)
print(df_bar)
print(df_bar1)
print(df_bar2)
print(df_bar3)
print(df_bar4)

              AB CD         F
sample1    AAAA1  C -1.395423
sample2    AAAA1  C  0.315828
sample3    AAAA1  C -1.894856
sample4    AAAA1  D -0.207946
sample5    AAAA1  D -0.142276
sample6   BBBBB2  D  0.552056
sample7   BBBBB2  G -0.176148
sample8   BBBBB2  G  0.169970
sample9   BBBBB2  G -1.398254
sample10  BBBBB2  G -1.422476
            Gene1     Gene2     Gene3     Gene4
sample1  0.170749  0.369818 -0.406722  0.831243
sample2  1.356510 -1.586601  0.360872 -1.808163
sample3 -0.142937 -0.402044 -1.372085 -0.006017
sample4 -1.960390 -2.046558 -0.034469  1.413408
sample5  0.462673  0.151571  0.420617  0.403127
           Scatter
sample1   2.845265
sample2   3.092774
sample3   9.248764
sample4   5.926675
sample5   3.709297
sample6   3.385716
sample7   6.177794
sample8   2.923336
sample9   4.870179
sample10  1.979889
              TMB1      TMB2
sample1   8.773112  1.653713
sample2   7.586741  7.328379
sample3   5.791634  4.394594
sample4   2.631353  2.556334
sample5   2.043007  6.890823
sample6   6.906687  7.230015
sample7   0.072210  0.240389
sample8   4.331631  8.055023
sample9   6.405677  8.152781
sample10  7.203812  3.805164
              T1-A      T1-B
sample1   6.962736  7.163614
sample2   0.599616  2.111658
sample3   1.991862  7.737146
sample4   5.307607  2.107476
sample5   5.391606  5.642582
sample6   9.262886  0.798720
sample7   0.149710  9.832642
sample8   5.501970  0.465060
sample9   4.053081  3.171087
sample10  5.727097  6.287463
              T2-A      T2-B
sample1   9.541359  0.763995
sample2   0.407472  5.127191
sample3   3.910844  7.810868
sample4   9.638641  4.979790
sample5   6.229322  4.687453
sample6   4.682800  2.185468
sample7   0.316455  2.642579
sample8   3.042201  0.237428
sample9   7.226572  4.233363
sample10  6.226147  5.776414
              T3-A      T3-B
sample1   6.367489  1.073648
sample2   8.109799  7.400489
sample3   0.510005  8.161129
sample4   1.897186  5.820624
sample5   4.127147  9.799400
sample6        NaN  5.438595
sample7   2.478326  1.461691
sample8   8.589167  1.485746
sample9   8.187439  0.822651
sample10  1.070760  6.860972
                T4
sample1   5.839575
sample2   7.886911
sample3   8.504942
sample4   5.119021
sample5   1.859604
sample6   6.721850
sample7   0.545908
sample8        NaN
sample9   7.443164
sample10  1.677521

[10]:

plt.figure(figsize=(4, 8))
col_ha = HeatmapAnnotation(
            label=anno_label(df.AB, merge=True,rotation=15),
            AB=anno_simple(df.AB,add_text=True), axis=1,
            CD=anno_simple(df.CD, add_text=True,text_kws={'color':'black'}),
            Exp=anno_boxplot(df_box, cmap='turbo',grid=True),
            Scatter=anno_scatterplot(df_scatter,legend=True,grid=True),
            Line=anno_lineplot(df_bar2,linewidth=4,colors={'T2-B':'orangered','T2-A':'yellowgreen'},
                              marker='D',legend=True),  #colors=['orangered','yellowgreen']
            TMB_bar=anno_barplot(df_bar,legend=True,cmap='Set1'),
            Bar1=anno_barplot(df_bar1,legend=True,colors=['red','black']), #colors can be str, list, tuple or dict
            Bar4=anno_barplot(df_bar4,legend=True,cmap='turbo'),
            Bar2=anno_barplot(df_bar2,legend=True,colors={'T2-B':'orangered','T2-A':'yellowgreen'}),
            Bar3=anno_barplot(df_bar3,legend=True,cmap='Paired'),
            plot=True,legend=True,legend_gap=5,hgap=4,
            # legend_order=False,
            # legend_order=['AB','CD','Line','Bar1'],legend_width=20
)
col_ha.show_ticklabels(df.index.tolist(),fontdict={'color':'blue'},rotation=-30)
plt.show()

Starting plotting HeatmapAnnotations
Collecting annotation legends..

../_images/notebooks_advanced_usage_13_1.png

Change orentation to `down` and add extra space¶

[11]:

plt.figure(figsize=(4, 8))
row_ha = HeatmapAnnotation(
            TMB_bar=anno_barplot(df_bar,legend=True,cmap='Set1'),
            Bar1=anno_barplot(df_bar1,legend=True,cmap='Dark2'),
            Bar4=anno_barplot(df_bar4,legend=True,cmap='turbo'),
            Bar2=anno_barplot(df_bar2,legend=True,cmap='tab10'),
            Bar3=anno_barplot(df_bar3,legend=True,cmap='Paired'),
            Scatter=anno_scatterplot(df_scatter),
            Exp=anno_boxplot(df_box, cmap='turbo',legend=True,grid=True),
            CD=anno_simple(df.CD, colors={'C': 'red', 'D': 'gray', 'G': 'yellow'},
                           add_text=True,legend=True,text_kws={'color':'black'}),
            AB=anno_simple(df.AB,add_text=True,legend=True),
            label=anno_label(df.AB, merge=True,rotation=-15),
            plot=True,plot_legend=False,legend_hpad=13,axis=1,hgap=1
            )
row_ha.show_ticklabels(df.index.tolist(),fontdict={'color':'blue'},rotation=30)
plt.show()
# Here, we can use hgap (when axis=1) or wgap (when axis=0) to control the widh of height space between different annotations.

Starting plotting HeatmapAnnotations

../_images/notebooks_advanced_usage_15_1.png

Change orentation to the left¶

[12]:

plt.figure(figsize=(8, 4))
row_ha = HeatmapAnnotation(
            label=anno_label(df.AB, merge=True,rotation=15),
            AB=anno_simple(df.AB,add_text=True,legend=True,
                           #text_kws=dict(bbox={"pad":0},va='center',ha='center',rotation_mode='anchor')
                          ),
            CD=anno_simple(df.CD,add_text=True,legend=True),
            Exp=anno_boxplot(df_box, cmap='turbo',legend=True),
            Scatter=anno_scatterplot(df_scatter),
            TMB_bar=anno_barplot(df_bar,legend=True,cmap='Set1'),
            Bar1=anno_barplot(df_bar1,legend=True,cmap='Dark2'),
            Bar4=anno_barplot(df_bar4,legend=True,cmap='turbo'),
            Bar2=anno_barplot(df_bar2,legend=True,cmap='tab10'),
            Bar3=anno_barplot(df_bar3,legend=True,cmap='Paired'),
            plot=True,legend=True,legend_gap=5,
            axis=0,legend_hpad=20,label_side='bottom',wgap=3,
            )
row_ha.show_ticklabels(df.index.tolist(),fontdict={'color':'blue'},rotation=0)
plt.show()

Starting plotting HeatmapAnnotations
Collecting annotation legends..
Incresing ncol
Incresing ncol
Incresing ncol
More than 3 cols is not supported
Legend too long, generating a new column..

../_images/notebooks_advanced_usage_17_1.png

Change orentation to the right¶

[13]:

plt.figure(figsize=(8, 4))
row_ha = HeatmapAnnotation(
            TMB_bar=anno_barplot(df_bar,legend=True,cmap='Set1'),
            Bar1=anno_barplot(df_bar1,legend=True,cmap='Dark2'),
            Bar4=anno_barplot(df_bar4,legend=True,cmap='turbo'),
            Bar2=anno_barplot(df_bar2,legend=True,cmap='tab10'),
            Bar3=anno_barplot(df_bar3,legend=True,cmap='Paired'),
            Scatter=anno_scatterplot(df_scatter,grid=True),
            Exp=anno_boxplot(df_box, cmap='turbo',legend=True,grid=True),
            CD=anno_simple(df.CD, colors={'C': 'red', 'D': 'gray', 'G': 'green'},
                           add_text=True,legend=True,
                           text_kws={'rotation':-90}),
            AB=anno_simple(df.AB,add_text=True,legend=True,
                           text_kws={'rotation':-90,'color':'black'}),
            label=anno_label(df.AB, merge=True,rotation=15),
            plot=True,legend=True,legend_hpad=13,legend_gap=5,axis=0,wgap=3,
            )
row_ha.show_ticklabels(df.index.tolist(),fontdict={'color':'black'},rotation=0)
plt.show()

Starting plotting HeatmapAnnotations
Collecting annotation legends..
Incresing ncol
Incresing ncol
Incresing ncol
More than 3 cols is not supported
Legend too long, generating a new column..

../_images/notebooks_advanced_usage_19_1.png

Changing orientation using parameter `orientation`¶

By Default, if there is no anno_label in the annotation, the oriention would be determined by parameter orientation.

[14]:

plt.figure(figsize=(8, 4))
col_ha = HeatmapAnnotation(
            AB=anno_simple(df.AB,add_text=True,legend=True),
            CD=anno_simple(df.CD,add_text=True,legend=True),
            Exp=anno_boxplot(df_box, cmap='turbo',legend=True),
            Scatter=anno_scatterplot(df_scatter,grid=True),
            TMB_bar=anno_barplot(df_bar,legend=True,cmap='Set1'),
            Bar1=anno_barplot(df_bar1,legend=True,cmap='Dark2'),
            Bar4=anno_barplot(df_bar4,legend=True,cmap='turbo'),
            Bar2=anno_barplot(df_bar2,legend=True,cmap='tab10'),
            Bar3=anno_barplot(df_bar3,legend=True,cmap='Paired'),
            plot=True,legend=True,axis=0,
            legend_gap=5,orientation='left',
            )
plt.show()

Starting plotting HeatmapAnnotations
Collecting annotation legends..
Incresing ncol
Incresing ncol
Incresing ncol
More than 3 cols is not supported
Legend too long, generating a new column..

../_images/notebooks_advanced_usage_21_1.png

[15]:

plt.figure(figsize=(8, 4))
col_ha = HeatmapAnnotation(
            AB=anno_simple(df.AB,add_text=True,legend=True,
                          text_kws={'rotation':-90,'fontsize':14,'color':'black'}),
            CD=anno_simple(df.CD,add_text=True,legend=True,
                          text_kws={'rotation':-90,'fontsize':14,'color':'white'}),
            Exp=anno_boxplot(df_box, cmap='turbo',legend=True),
            Scatter=anno_scatterplot(df_scatter),
            TMB_bar=anno_barplot(df_bar,legend=True,cmap='Set1'),
            Bar1=anno_barplot(df_bar1,legend=True,cmap='Dark2'),
            Bar4=anno_barplot(df_bar4,legend=True,cmap='turbo'),
            Bar2=anno_barplot(df_bar2,legend=True,cmap='tab10'),
            Bar3=anno_barplot(df_bar3,legend=True,cmap='Paired'),
            plot=True,legend=True,axis=0,wgap=3,
            legend_gap=5,orientation='right',
            )
plt.show()

Starting plotting HeatmapAnnotations
Collecting annotation legends..
Incresing ncol
Incresing ncol
Incresing ncol
More than 3 cols is not supported
Legend too long, generating a new column..

../_images/notebooks_advanced_usage_22_1.png

Add multiple heatmap annotations using `for` loop¶

Typically, we can create a heatmap annotatin using the following code:

col_ha = HeatmapAnnotation(
                        Group=anno_simple(df_cols.hypomethylated_samples,colors=sample_group_color_dict,legend=True),
                        CellType=anno_simple(df_cols.CellType,colors=ct_color_dict,legend=ct_legend),
                        M1=anno_simple(df_cols['M1'],cmap='jet',legend=lgd,vmax=1,vmin=0,legend_kws={'label':'M1'}),
                        verbose=0,label_side='right',label_kws={'horizontalalignment':'left'})

But what if we have many annotations, for example:

col_ha = HeatmapAnnotation(
                        Group=anno_simple(df_cols.hypomethylated_samples,colors=sample_group_color_dict,legend=True),
                        CellType=anno_simple(df_cols.CellType,colors=ct_color_dict,legend=ct_legend),
                        M1=anno_simple(df_cols['M1'],cmap='jet',legend=lgd,vmax=1,vmin=0,legend_kws={'label':'M1'}),
                        M2=anno_simple(df_cols['M2'],cmap='jet',legend=lgd,vmax=1,vmin=0,legend_kws={'label':'M2'}),
                        M3=anno_simple(df_cols['M3'],cmap='jet',legend=lgd,vmax=1,vmin=0,legend_kws={'label':'M3'}),
                        .....
                        verbose=0,label_side='right',label_kws={'horizontalalignment':'left'})

In this case, we can create an dict including the name and annotation as keys and values:

col_ha_dict={
                'Group':anno_simple(df_cols.hypomethylated_samples,colors=sample_group_color_dict,legend=True),
                'CellType':anno_simple(df_cols.CellType,colors=ct_color_dict,legend=ct_legend)
                }


for col in sample_cols:
    col_ha_dict[col]=anno_simple(df_cols[col],cmap='jet',legend=lgd,vmax=1,vmin=0,legend_kws={'label':col})

col_ha = HeatmapAnnotation(**col_ha_dict,
                           verbose=0,label_side='right',label_kws={'horizontalalignment':'left'})

Cluster between groups and cluster within groups¶

Similar to cluster_between_groups and cluster_within_groups in R (https://jokergoo.github.io/2021/03/05/cluster-groups-in-complexheatmap/)

clsuter within groups: `col_split=*, col_cluster=True`¶

[16]:

df['Groups']=['G1']+['G2']+['G3']*5+['G4']+['G5']*2
col_ha = HeatmapAnnotation(
            Groups=anno_simple(df.Groups,add_text=True,text_kws={'color':'black'}),
            AB=anno_simple(df.AB,add_text=True),axis=1,
            Exp=anno_boxplot(df_box, cmap='turbo'),
            verbose=0) #verbose=0 will turn off the log.


plt.figure(figsize=(6, 8))
cm = ClusterMapPlotter(
        data=df_heatmap, top_annotation=col_ha,
        col_split=df.Groups,col_split_gap=2,
        col_cluster=True,row_cluster=True,col_dendrogram=True,
        label='values',show_rownames=True,show_colnames=True,
        tree_kws={'col_cmap': 'Set1'},verbose=0,legend_gap=7,
        annot=True,fmt='.1g',linewidths=0.05,linecolor='gold',cmap='RdYlBu_r',
        xticklabels_kws={'labelrotation':-45,'labelcolor':'blue'},
        ylabel='Features',
        legend_order=['AB','Groups','Exp','values'] #change legend order
)
plt.show()

../_images/notebooks_advanced_usage_26_0.png

cluster_between_groups: `col_split=*, col_split_order="cluster_between_groups",col_cluster=False`¶

[17]:

plt.figure(figsize=(8, 10))
cm = ClusterMapPlotter(
        data=df_heatmap, top_annotation=col_ha,
        col_split=df.Groups, col_split_order="cluster_between_groups",
        col_split_gap=2,col_cluster=False,
        row_cluster=True,col_dendrogram=True,row_dendrogram_size=35,col_dendrogram_size=25,
        row_split=2,row_split_gap=1,row_dendrogram=True,
        label='values',show_rownames=True,show_colnames=True,bezier=True,dotsize=8,
        tree_kws={'colors':'blue','row_cmap':'Set1','col_cmap':'Paired'},
        verbose=0,legend_gap=7,
        linewidths=0.05,linecolor='gold',cmap='RdYlBu_r',
        xticklabels_kws={'labelrotation':-45,'labelcolor':'blue'},
        ylabel='Features')
plt.show()

../_images/notebooks_advanced_usage_28_0.png

cluster_within_groups && cluster_between_groups: `col_split=*, col_split_order="cluster_between_groups",col_cluster=True`¶

[18]:

plt.figure(figsize=(6, 8))
cm = ClusterMapPlotter(
        data=df_heatmap, top_annotation=col_ha,
        col_split=df.loc[:,['AB','Groups']], col_split_order="cluster_between_groups",
        col_split_gap=2,col_cluster=True,row_split_gap=1.5,
        row_split=3,#row_split_order='cluster_between_groups',
        row_cluster=True,col_dendrogram=True,row_dendrogram=True,
        label='values',show_rownames=True,show_colnames=True,
        tree_kws={'colors':'blue'},verbose=0,legend_gap=7,
        annot=True,fmt='.1g',linewidths=0.05,linecolor='gold',cmap='RdYlBu_r',
        xticklabels_kws={'labelrotation':-45,'labelcolor':'blue'},
        ylabel='Features')
plt.show()

../_images/notebooks_advanced_usage_30_0.png

[19]:

# `label_kws` in `HeatmapAnnotation` control the heatmap annotaiton labels
col_ha = HeatmapAnnotation(
                Groups=anno_simple(df.Groups,add_text=True,text_kws={'color':'black'}),
                AB=anno_simple(df.AB,add_text=True),axis=1,
                Exp=anno_boxplot(df_box, cmap='turbo',grid=True),
                verbose=0,label_side='right'
)

# `xticklabels_kws` and `yticklabels_kws` control the ticklabels for the heatmap.
plt.figure(figsize=(6, 8))
cm = ClusterMapPlotter(data=df_heatmap, top_annotation=col_ha,
                       col_split=df.Groups,col_split_order=['G2','G1','G5','G4','G3'],
                       col_split_gap=4.5,col_cluster=True,
                       row_cluster=True,col_dendrogram=True,
                       label='values',show_rownames=True,show_colnames=True,
                       row_names_side='left',
                       tree_kws={'col_cmap':'Set1'},verbose=0,legend_gap=7,
                       linewidths=0.05,linecolor='gold',cmap='RdYlBu_r',
                       xticklabels_kws=dict(labelrotation=-45,labelcolor='purple',labelsize=14),
                       #more parameters for [x/y]_ticklabels_kws, see: matplotlib.axes.Axes.tick_params or ?ax.tick_params
                       xlabel='Samples',ylabel="Features",
                       xlabel_kws=dict(color='white',fontsize=14),
                       ylabel_kws=dict(color='blue',fontsize=14,labelpad=45), #increace labelpad manually using labelpad (points)
                       xlabel_bbox_kws=dict(facecolor='black'),
                       ylabel_bbox_kws=dict(facecolor='chocolate',edgecolor='red'),
                      )
plt.savefig("test.pdf",bbox_inches='tight')
plt.show()

../_images/notebooks_advanced_usage_31_0.png

Custom annotation¶

[20]:

annot=df_heatmap.applymap(lambda x:'∗∗∗' if x >= 2 else '∗∗' if x >=1 else '∗' if x >0 else '')
# To make asterisk located at center in vertical, use ∗ ASTERISK OPERATOR. instead of normal *; see: https://unicode-explorer.com/c/2217
plt.figure(figsize=(5, 6.5))
cm = ClusterMapPlotter(
        data=df_heatmap, top_annotation=col_ha,
        annot=annot,fmt=None,annot_kws={'color':'white','fontname':'Courier'},
        col_split=df.Groups, col_split_order="cluster_between_groups",
        col_cluster=True,row_cluster=True,
        label='values',
        tree_kws={'col_cmap': 'Set1'},verbose=0,legend_gap=7,
        linewidths=0.05,linecolor='white',cmap='RdYlBu_r',
        xticklabels_kws={'labelrotation':-45,'labelcolor':'blue'})
plt.show()

../_images/notebooks_advanced_usage_33_0.png

Custom linkage¶

[21]:

import fastcluster
# custom column linkage
linkage = fastcluster.linkage(df_heatmap.T.apply(lambda x:x.fillna(x.median()),axis=1), method='average', metric='canberra')
print("df_heatmap shape:",df_heatmap.shape,"\nlinkage shape:",linkage.shape,"\n",linkage)

plt.figure(figsize=(4, 6))
cm = ClusterMapPlotter(
        data=df_heatmap, top_annotation=col_ha,z_score=0,
        col_cluster=True,row_cluster=True,show_rownames=True,show_colnames=True,
        label='values',col_dendrogram_kws=dict(linkage=linkage),col_dendrogram=True,
        tree_kws={'col_cmap': 'Set1'},verbose=0,legend_gap=7,
        linewidths=0.05,linecolor='white',cmap='RdYlBu_r',
        xticklabels_kws={'labelrotation':-45,'labelcolor':'blue'})
plt.show()

df_heatmap shape: (30, 10)
linkage shape: (9, 4)
 [[ 1.          6.         16.39057782  2.        ]
 [ 3.          8.         16.59636466  2.        ]
 [ 7.         11.         18.67771195  3.        ]
 [ 4.         10.         19.30874384  3.        ]
 [ 5.          9.         19.51326053  2.        ]
 [ 0.          2.         20.63068239  2.        ]
 [12.         13.         20.65079744  6.        ]
 [15.         16.         22.1959461   8.        ]
 [14.         17.         23.33605841 10.        ]]

../_images/notebooks_advanced_usage_35_1.png

[22]:

df['Groups']=['G1']+['G2']+['G3']*5+['G4']+['G5']*2
plt.figure(figsize=(4, 6))
cm = ClusterMapPlotter(
        data=df_heatmap, top_annotation=col_ha,
        col_cluster=True,row_cluster=True,show_rownames=True,show_colnames=True,
        row_split=2,row_split_gap=3,row_dendrogram=True,
        label='values',col_dendrogram_kws=dict(linkage=linkage),col_dendrogram=True,
        tree_kws={'col_cmap': 'Set1','row_cmap':'Dark2'},verbose=0,legend_gap=7,
        linewidths=0.05,linecolor='white',cmap='RdYlBu_r',
        xticklabels_kws={'labelrotation':-45,'labelcolor':'blue'})
plt.show()

../_images/notebooks_advanced_usage_36_0.png

Image annotation¶

[23]:

df = pd.DataFrame(['AAAA1'] * 5 + ['BBBBB2'] * 5, columns=['AB'])
df['CD'] = ['C'] * 3 + ['D'] * 3 + ['G'] * 4
df['F'] = np.random.normal(0, 1, 10)
df.index = ['sample' + str(i) for i in range(1, df.shape[0] + 1)]
df_box = pd.DataFrame(np.random.randn(10, 4), columns=['Gene' + str(i) for i in range(1, 5)])
df_box.index = ['sample' + str(i) for i in range(1, df_box.shape[0] + 1)]
df_bar = pd.DataFrame(np.random.uniform(0, 10, (10, 2)), columns=['TMB1', 'TMB2'])
df_bar.index = ['sample' + str(i) for i in range(1, df_box.shape[0] + 1)]
df_scatter = pd.DataFrame(np.random.uniform(0, 10, 10), columns=['Scatter'])
df_scatter.index = ['sample' + str(i) for i in range(1, df_box.shape[0] + 1)]
df_bar1 = pd.DataFrame(np.random.uniform(0, 10, (10, 2)), columns=['T1-A', 'T1-B'])
df_bar1.index = ['sample' + str(i) for i in range(1, df_box.shape[0] + 1)]
df_bar2 = pd.DataFrame(np.random.uniform(0, 10, (10, 2)), columns=['T2-A', 'T2-B'])
df_bar2.index = ['sample' + str(i) for i in range(1, df_box.shape[0] + 1)]
df_bar3 = pd.DataFrame(np.random.uniform(0, 10, (10, 2)), columns=['T3-A', 'T3-B'])
df_bar3.index = ['sample' + str(i) for i in range(1, df_box.shape[0] + 1)]
df_bar3.iloc[5,0]=np.nan
df_bar4 = pd.DataFrame(np.random.uniform(0, 10, (10, 1)), columns=['T4'])
df_bar4.index = ['sample' + str(i) for i in range(1, df_box.shape[0] + 1)]
df_bar4.iloc[7,0]=np.nan
df_img = pd.DataFrame(['https://upload.wikimedia.org/wikipedia/commons/thumb/c/c3/Python-logo-notext.svg/121px-Python-logo-notext.svg.png',
                       'https://motifcollections.aertslab.org/v10nr_clust/logos/metacluster_135.7.png',
                       'https://cdn3.iconfinder.com/data/icons/family-member-flat-happy-family-day/512/Brother-512.png',
                       'https://cdn3.iconfinder.com/data/icons/family-member-flat-happy-family-day/512/Sister-512.png',
                       'https://img.freepik.com/free-vector/sticker-design-with-cute-mouse-isolated_1308-59360.jpg',
                       'https://motifcollections.aertslab.org/v10nr_clust/logos/metacluster_131.8.png',
                       'https://img.freepik.com/premium-vector/vector-illustration-gorilla-isolated-white-background-cartoon-style_1151-66575.jpg',
                      "2.png",'1.jpeg',
                      'https://upload.wikimedia.org/wikipedia/commons/thumb/c/c3/Python-logo-notext.svg/121px-Python-logo-notext.svg.png'], columns=['path'])
df_img.index = ['sample' + str(i) for i in range(1, df_box.shape[0] + 1)]

[24]:

plt.figure(figsize=(16, 4))
col_ha = HeatmapAnnotation(
            label=anno_label(df.AB, merge=True,rotation=15),
            AB=anno_simple(df.AB,add_text=True,legend=True), axis=1,
            CD=anno_simple(df.CD, add_text=True,legend=True,text_kws={'color':'black'}),
            Exp=anno_boxplot(df_box, cmap='turbo',legend=True),
            Scatter=anno_scatterplot(df_scatter),
            Bar1=anno_barplot(df_bar1,legend=True,cmap='Dark2'),
            Bar4=anno_barplot(df_bar4,legend=True,cmap='turbo'),
            Img=anno_img(df_img.path,border_width=5,border_color=255,height=15),
            plot=True,legend=True,legend_gap=5,hgap=0.5)
col_ha.show_ticklabels(df.index.tolist(),fontdict={'color':'blue'},rotation=-30)
plt.show()

Starting plotting HeatmapAnnotations
Collecting annotation legends..

../_images/notebooks_advanced_usage_39_1.png

How to force display all row/col ticklabels?¶

When the height or width is not big enough to display all xticklabels and yticklabels, some ticklabels will be hidden to avoid overlapping. For example:

[25]:

plt.figure(figsize=(3.5, 5))
cm = ClusterMapPlotter(
        data=df_heatmap,
        col_cluster=True,row_cluster=True,
        col_split=df.AB,row_split=2,
        col_split_gap=0.5,row_split_gap=0.8,
        label='values',row_dendrogram=True,
        show_rownames=True,show_colnames=True,row_names_side='right',
        tree_kws={'row_cmap': 'Set1'},verbose=0,legend_gap=5,
        cmap='meth2',xticklabels_kws={'labelrotation':-90,'labelcolor':'blue'})
plt.show()

../_images/notebooks_advanced_usage_41_0.png

To force display all ticklabels no matter whether the height or width is big enough, set parameters xticklabels/yticklabels to True:

[26]:

plt.figure(figsize=(3.5, 5))
cm = ClusterMapPlotter(
        data=df_heatmap,
        col_cluster=True,row_cluster=True,
        col_split=df.AB,row_split=2,
        col_split_gap=0.5,row_split_gap=0.8,
        label='values',row_dendrogram=True,
        show_rownames=True,show_colnames=True,
        row_names_side='right',yticklabels=True,
        tree_kws={'row_cmap': 'Set1'},verbose=0,legend_gap=5,
        cmap='meth2',xticklabels_kws={'labelrotation':-90,'labelcolor':'blue'})
plt.show()

../_images/notebooks_advanced_usage_43_0.png