Vocabulary/OONMFvis.py at master · Altius/Vocabulary · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
ClusterMode = False

import numpy as np
import pandas as pd
if (ClusterMode):
	import matplotlib
	matplotlib.use('Agg')

import matplotlib.pyplot as plt
import seaborn as sns
import matplotlib.gridspec as gridspec


def return_expected_matrix(Gcut, maxcolor, secondmaxcolor, secondmaxcolor_vals, strings_of_labels,  rownormed=False):

    bincount_matrix_cutoff = []
    for comp in range(16):
        bincut_row = []
        maincut = maxcolor[Gcut] == comp
        for j in range(16):
            seccut = (secondmaxcolor[Gcut] == j) *  (secondmaxcolor_vals[Gcut]>0)
            if comp == j:
                seccut = secondmaxcolor_vals[Gcut]==0
            N_in_square_cut = maincut * seccut
            N_in_square = np.sum(N_in_square_cut)
            bincut_row.append(N_in_square)
        bincount_matrix_cutoff.append(bincut_row)
    bincount_matrix_cutoff = np.array(bincount_matrix_cutoff)
    if rownormed:
        bincount_matrix_cutoff_rownormed = (bincount_matrix_cutoff.T / np.sum(bincount_matrix_cutoff, axis=1)).T
        bincount_matrix_cutoff_rownormedPD = pd.DataFrame(bincount_matrix_cutoff_rownormed, columns=strings_of_labels)
        bincount_matrix_cutoff_rownormedPD['domcomp'] = strings_of_labels
        bincount_matrix_cutoff_rownormedPD.set_index('domcomp', inplace=True)

    bincount_matrix_cutoffPD = pd.DataFrame(bincount_matrix_cutoff, columns=strings_of_labels)
    bincount_matrix_cutoffPD['domcomp'] = strings_of_labels
    bincount_matrix_cutoffPD.set_index('domcomp', inplace=True)

        #return both  counts and row-normed version
    if rownormed:
        return [bincount_matrix_cutoffPD, bincount_matrix_cutoff_rownormedPD]

    return bincount_matrix_cutoffPD

# not the prettiest function, but
def big_grid_plot(Gcut, maxcolor, secondmaxcolor, maxcolor_vals, secondmaxcolor_vals, WSO, color_scheme, strings_of_labels, plt_title = '', do_row_normalized_shading=True, do_extra_bargraphs=True, logOOEmode=False, adjust_significance=False, ExpectedMatrix = [], pmatrix = [] ):
    maxcolor_bincount = np.bincount(maxcolor[Gcut], minlength=16)
    seccolor_bincount = np.bincount(secondmaxcolor[Gcut], minlength=16)

    seccut = secondmaxcolor_vals>0
    # possibly need to revise this - by default, the expectation value is just all sites
    maxcolor_bincount_OOE = np.bincount(maxcolor[Gcut], minlength=16) /  np.bincount(maxcolor, minlength=16) * (len(maxcolor) / len(maxcolor[Gcut]))

    seccolor_bincount_OOE =  np.bincount(secondmaxcolor[Gcut*seccut], minlength=16) /  np.bincount(secondmaxcolor[seccut], minlength=16) * (len(secondmaxcolor[seccut]) / len(secondmaxcolor[Gcut* seccut]))


    if do_row_normalized_shading:
        [bincount_matrix_cutoff_G_PD, bincount_matrix_cutoff_G_rownormedPD] = return_expected_matrix(Gcut,maxcolor, secondmaxcolor, secondmaxcolor_vals, strings_of_labels, rownormed=True)
    else:
        bincount_matrix_cutoff_G_PD = return_expected_matrix(Gcut,maxcolor, secondmaxcolor,secondmaxcolor_vals, strings_of_labels)


    gs1 = gridspec.GridSpec(5,5)
    fig = plt.figure(figsize = (24,22))
    plt.clf()
    f = plt.subplot2grid((5,5), (0,1), colspan=3, rowspan=3)


    if (do_row_normalized_shading):
        kox= sns.heatmap(bincount_matrix_cutoff_G_rownormedPD[strings_of_labels[WSO]].reindex(strings_of_labels[WSO]), annot=bincount_matrix_cutoff_G_PD[strings_of_labels[WSO]].reindex(strings_of_labels[WSO]).values.astype(int), cmap='binary', vmax=0.5, fmt='d', cbar_kws={'fraction':0.046, 'pad':0.04} )
        kox.collections[0].colorbar.set_label('secondary fraction', fontsize=25)
    elif logOOEmode:
        if len(ExpectedMatrix) < 1:
            print('error, must pass return matrix')
            plt.close()
            return
        group_rat = bincount_matrix_cutoff_G_PD / ExpectedMatrix
        group_fix = ExpectedMatrix.sum().sum() / bincount_matrix_cutoff_G_PD.sum().sum()

        thing_to_heatmap = np.log10(group_fix*group_rat[strings_of_labels[WSO]].reindex(strings_of_labels[WSO])+1e-5) / np.log10(2) #using log2

        if (adjust_significance):
            temp_annotations = thing_to_heatmap.values
            annotations = []

            for i in range(len(temp_annotations)):
                irow = []
                for j in range(len(temp_annotations)):
                    cellvalue = '{:6.2f}'.format(temp_annotations[i][j]).strip(' ')
                    if pmatrix[i][j]:
                        irow.append(cellvalue+'*')
                    else:
                        irow.append(cellvalue)
                annotations.append(irow)
            annotations = np.array(annotations)

            #annotations = temp_annotations
            annotationsPD = pd.DataFrame(annotations, index =strings_of_labels[WSO], columns = strings_of_labels[WSO] )
            print(annotations.shape)
            kox= sns.heatmap(thing_to_heatmap, annot=annotationsPD, cmap='RdYlGn', vmax=8, vmin=-8, fmt='s', cbar_kws={'fraction':0.046, 'pad':0.04}, annot_kws={'fontsize':12}, mask =  thing_to_heatmap<-10 )

        else:
            kox= sns.heatmap(thing_to_heatmap, annot=True, cmap='RdYlGn', vmax=8, vmin=-8, fmt='6.2f', cbar_kws={'fraction':0.046, 'pad':0.04}, mask =  thing_to_heatmap<-10 )


        kox.collections[0].colorbar.set_label('log2 O/E', fontsize=25)


    ax =plt.gca()
    kox.collections[0].colorbar.ax.tick_params(labelsize=25)
    kox.set_aspect("equal")
    the_fontsize=24
    ax.set_ylabel('primary component',fontsize=20)
    ax.set_xlabel('secondary component', fontsize=20)
    ax.get_xticklabels()
    for n, i in enumerate(ax.get_xticklabels()):
        i.set_color(color_scheme[n])
        i.set_fontsize(the_fontsize)
    for n, i in enumerate(ax.get_yticklabels()):
        i.set_fontsize(the_fontsize)
        i.set_color(color_scheme[n])
    plt.title(plt_title, fontsize=40)


    if (do_extra_bargraphs):

        #left plot for O/E for main comp

        leftplot = plt.subplot2grid((5,5), (0,0), rowspan=3, position=[-0.05, 0.395, 0.1, 0.525])

        leftplot.barh(['C'+str(i+1) for i in range(16)], maxcolor_bincount_OOE[WSO][::-1], color=color_scheme[::-1], tick_label=strings_of_labels[WSO][::-1])
        leftplot.plot([1, 1], [-1, 16], '--k')
        leftplot.plot([2, 2], [-1, 16], '--k')

        leftplot_ax = plt.gca()
        leftplot_ax.set_xticks([1 , 2])
        leftplot_ax.set_yticklabels([])
        leftplot_ax.set_xticklabels(['1', '2'])
        leftplot_ax.get_yaxis().set_visible(False)
        for n, i in enumerate(leftplot_ax.get_xticklabels()):
            i.set_fontsize(20)
        leftplot_ax.invert_xaxis()
        plt.xlabel('O/E', fontsize=24)
        plt.box(False)

        #bottom plot for secondary O/E
        botplot = plt.subplot2grid((5,5), (4,0), colspan=3, position=[0.255, 0.1, 0.485, 0.1])
        botplot.bar(['C'+str(i+1) for i in range(16)], seccolor_bincount_OOE[WSO], color=color_scheme, tick_label=strings_of_labels[WSO])
        botplot.plot( [-1, 16], [1, 1], '--k')
        botplot.plot( [-1, 16], [2, 2], '--k')
        botplot_ax = plt.gca()
        botplot_ax.set_yticks([1 , 2])
        botplot_ax.set_xticklabels([])
        botplot_ax.set_yticklabels(['1', '2'])
        botplot_ax.get_xaxis().set_visible(False)
        for n, i in enumerate(botplot_ax.get_yticklabels()):
            i.set_fontsize(20)
        botplot_ax.invert_yaxis()
        plt.ylabel('O/E', fontsize=24)
        plt.box(False)

    plt.savefig('test.pdf', bbox_inches='tight', transparent=True)

    plt.show()

    plt.close()