forked from rueuntal/TL
-
Notifications
You must be signed in to change notification settings - Fork 0
/
TL_plot.py
242 lines (211 loc) · 12.2 KB
/
TL_plot.py
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
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
## Test the plotting function
from __future__ import division
import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
import TL_functions as tl
import numpy as np
from scipy import stats
import random
# Figure 1 - visual representation using three studies
study_list = ['1_1', '10_1', '52_11']
fig = plt.figure(figsize = (10.5, 7))
iplot = 1
for feas_type in ['partition', 'composition']:
for study in study_list:
ax = plt.subplot(2, 3, iplot)
if iplot == 1 or iplot == 4: legend = True
else: legend = False
tl.plot_emp_vs_sim(study, feas_type = feas_type, ax = ax, legend = legend)
iplot += 1
plt.subplots_adjust(wspace = 0.29, hspace = 0.29)
plt.savefig('Fig1.pdf', dpi = 600)
# Figure 2 - compare the full distribution of empirical TLs and those from the feasible sets
study_info = tl.get_study_info('study_taxon_type.txt')
tl_pars_par = tl.get_tl_par_file('out_files/TL_form_partition.txt')
var_par = tl.get_var_sample_file('out_files/taylor_QN_var_predicted_partition_1000_full.txt')
var_comp = tl.get_var_sample_file('out_files/taylor_QN_var_predicted_composition_1000_full.txt')
par_quad = tl.get_val_ind_sample_file('out_files/TL_quad_p_partition.txt')
comp_quad = tl.get_val_ind_sample_file('out_files/TL_quad_p_composition.txt')
b_obs, b_par, b_comp, b_type = [], [], [], []
p_obs, p_par, p_comp = [], [], []
pcurv_obs, pcurv_par, pcurv_comp = [], [], []
r2_obs, r2_par, r2_comp = [], [], []
for study in np.unique(var_par['study']):
b_obs.append((tl_pars_par['b_obs'][tl_pars_par['study'] == study])[0])
p_obs.append((tl_pars_par['p_obs'][tl_pars_par['study'] == study])[0])
r2_obs.append((tl_pars_par['R2_obs'][tl_pars_par['study'] == study])[0])
pcurv_obs.append((par_quad['emp_val'][par_quad['study'] == study])[0])
b_type.append((study_info['type'][study_info['study'] == study])[0])
sample_par = var_par[var_par['study'] == study]
sample_comp = var_comp[var_comp['study'] == study]
for i in xrange(1000):
sample_par_i = [sample_par[x][i + 5] for x in xrange(len(sample_par))]
mean_par = [sample_par['mean'][p] for p in xrange(len(sample_par)) if sample_par_i[p] != 0]
sample_par_i = [sample_par_i[p] for p in xrange(len(sample_par_i)) if sample_par_i[p] != 0]
b_par_i, inter_par_i, r_par_i, p_par_i, std_err_par_i = stats.linregress(np.log(mean_par), np.log(sample_par_i))
b_par.append(b_par_i)
p_par.append(p_par_i)
r2_par.append(r_par_i ** 2)
pcurv_par.append(par_quad[par_quad['study'] == study][0][i + 2])
sample_comp_i = [sample_comp[y][i + 5] for y in xrange(len(sample_comp))]
mean_comp = [sample_comp['mean'][q] for q in xrange(len(sample_comp)) if sample_comp_i[q] != 0]
sample_comp_i = [sample_comp_i[q] for q in xrange(len(sample_comp_i)) if sample_comp_i[q] != 0]
b_comp_i, inter_comp_i, r_comp_i, p_comp_i, std_err_comp_i = stats.linregress(np.log(mean_comp), np.log(sample_comp_i))
b_comp.append(b_comp_i)
p_comp.append(p_comp_i)
r2_comp.append(r_comp_i ** 2)
pcurv_comp.append(comp_quad[comp_quad['study'] == study][0][i + 2])
fig = plt.figure(figsize = (7, 7))
ax_p = plt.subplot(221)
tl.plot_dens_par_comp(p_obs, p_par, p_comp, ax = ax_p, legend = True, loc = 1, vline = 0.05, xlim = [0, 0.2])
ax_p.annotate('(A)', xy = (0.05, 0.92), xycoords = 'axes fraction', fontsize = 10)
plt.xlabel('p-value for b', fontsize = 8)
plt.ylabel('Density', fontsize = 8)
ax_curv = plt.subplot(222)
tl.plot_dens_par_comp(pcurv_obs, pcurv_par, pcurv_comp, ax = ax_curv, vline = 0.05, xlim = [0, 1])
ax_curv.annotate('(B)', xy = (0.05, 0.92), xycoords = 'axes fraction', fontsize = 10)
plt.xlabel('p-value for quadratic term', fontsize = 8)
plt.ylabel('Density', fontsize = 8)
ax_r2 = plt.subplot(223)
tl.plot_dens_par_comp(r2_obs, r2_par, r2_comp, ax = ax_r2, xlim = [0, 1])
ax_r2.annotate('(C)', xy = (0.05, 0.92), xycoords = 'axes fraction', fontsize = 10)
plt.xlabel(r'$R^2$', fontsize = 8)
plt.ylabel('Density', fontsize = 8)
ax_b = plt.subplot(224)
tl.plot_dens_par_comp(b_obs, b_par, b_comp, ax = ax_b, xlim = [0, 4])
ax_b.annotate('(D)', xy = (0.05, 0.92), xycoords = 'axes fraction', fontsize = 10)
plt.xlabel('Exponent b', fontsize = 8)
plt.ylabel('Density', fontsize = 8)
plt.savefig('Fig2.pdf', dpi = 600)
# Figure 3 - compare empirical versus feasible set variance and b
tl_pars_par = tl.get_tl_par_file('out_files/TL_form_partition.txt')
study_sig = tl_pars_par['study']
var_par = tl.get_var_sample_file('out_files/taylor_QN_var_predicted_partition_1000_full.txt')
var_comp = tl.get_var_sample_file('out_files/taylor_QN_var_predicted_composition_1000_full.txt')
par_index = [i for i in range(len(var_par)) if var_par['study'][i] in study_sig]
var_par = var_par[par_index]
comp_index = [i for i in range(len(var_comp)) if var_comp['study'][i] in study_sig]
var_comp = var_comp[comp_index]
study_info = tl.get_study_info('study_taxon_type.txt')
# Here the values are relative to the emp value
expc_upper_par, expc_upper_comp, expc_par, expc_comp, expc_lower_par, expc_lower_comp = [], [], [], [], [], []
for i in xrange(len(var_par)):
expc_sample_par = [var_par[i][j] for j in xrange(5, 1005)]
expc_par.append(np.mean(expc_sample_par))
expc_lower_par.append(np.percentile(expc_sample_par, 2.5))
expc_upper_par.append(np.percentile(expc_sample_par, 97.5))
expc_sample_comp = [var_comp[i][j] for j in xrange(5, 1005)]
expc_comp.append(np.mean(expc_sample_comp))
expc_lower_comp.append(np.percentile(expc_sample_comp, 2.5) )
expc_upper_comp.append(np.percentile(expc_sample_comp, 97.5))
fig = plt.figure(figsize = (7, 7))
ax_par = plt.subplot(221)
tl.plot_obs_expc_new(var_par['var'], expc_par, expc_upper_par, expc_lower_par, 'partition', True, ax = ax_par)
plt.xlabel(r'Index for $s^2$', fontsize = 10)
plt.ylabel(r'$s_{partition}^2$ / $s_{empirical}^2$', fontsize = 12)
plt.title('Partitions')
ax_comp = plt.subplot(222)
tl.plot_obs_expc_new(var_comp['var'], expc_comp, expc_upper_comp, expc_lower_comp, 'composition', True, ax = ax_comp)
plt.xlabel(r'Index for $s^2$', fontsize = 10)
plt.ylabel(r'$s_{composition}^2$/ $s_{empirical}^2$', fontsize = 12)
plt.title('Compositions')
ax_b_par = plt.subplot(223)
tl_pars_par = tl.get_tl_par_file('out_files/TL_form_partition.txt')
tl.plot_obs_expc_new(tl_pars_par['b_obs'], tl_pars_par['b_expc'], tl_pars_par['b_upper'], \
tl_pars_par['b_lower'], 'partition', False, ax = ax_b_par)
plt.xlabel('Index for b', fontsize = 10)
plt.ylabel(r'$b_{partition}$ - $b_{empirical}$', fontsize = 12)
ax_b_comp = plt.subplot(224)
tl_pars_comp = tl.get_tl_par_file('out_files/TL_form_composition.txt')
tl.plot_obs_expc_new(tl_pars_comp['b_obs'], tl_pars_comp['b_expc'], tl_pars_comp['b_upper'], \
tl_pars_comp['b_lower'], 'composition', False, ax = ax_b_comp)
plt.xlabel('Index for b', fontsize = 10)
plt.ylabel(r'$b_{composition}$ - $b_{empirical}$', fontsize = 12)
plt.subplots_adjust(wspace = 0.29, hspace = 0.29)
plt.savefig('Fig3.pdf', dpi = 600)
# 10. Figure B1 - examples of empirical variance versus the full distribution from the feasible set
var_par = tl.get_var_sample_file('out_files/taylor_QN_var_predicted_partition_1000_full.txt')
var_comp = tl.get_var_sample_file('out_files/taylor_QN_var_predicted_composition_1000_full.txt')
tl_pars_par = tl.get_tl_par_file('out_files/TL_form_partition.txt')
random.seed(4)
qn_sets = random.sample(range(len(var_par)), 3)
fig = plt.figure(figsize = (10.5, 3.5))
# Plot the 3 Q-N pairs
for i, qn_pair in enumerate(qn_sets):
ax_qn = plt.subplot(1, 3, i + 1)
dat_row_par = list(var_par[qn_pair])
study_comp = var_comp[var_comp['study'] == dat_row_par[0]]
dat_row_comp = study_comp[(study_comp['Q'] == dat_row_par[1]) * (study_comp['N'] == dat_row_par[2])]
dat_row_comp = list(dat_row_comp[0])
if i == 0:
tl.plot_dens_par_comp_single_obs(dat_row_par[4], dat_row_par[5:], dat_row_comp[5:], \
ax = ax_qn, legend = True, loc = 2)
else: tl.plot_dens_par_comp_single_obs(dat_row_par[4], dat_row_par[5:], dat_row_comp[5:], ax = ax_qn)
ax_qn.annotate('Q = ' + str(dat_row_par[1]), xy = (0.7, 0.92), xycoords = 'axes fraction', fontsize = 10)
ax_qn.annotate('N = ' + str(dat_row_par[2]), xy = (0.7, 0.82), xycoords = 'axes fraction', fontsize = 10)
plt.xlabel('Variance', fontsize = 10)
plt.ylabel('Density', fontsize = 10)
plt.subplots_adjust(wspace = 0.29)
plt.savefig('FigB1.pdf', dpi = 600)
# Figure B2 - results from 4000 samples
study_info = tl.get_study_info('study_taxon_type.txt')
tl_pars_par = tl.get_tl_par_file('out_files/TL_form_partition_4000.txt')
var_par_1000 = tl.get_var_sample_file('out_files/taylor_QN_var_predicted_partition_1000_full.txt')
var_par = tl.get_var_sample_file('out_files/taylor_QN_var_predicted_partition_4000_full.txt', sample_size = 4000)
var_comp = tl.get_var_sample_file('out_files/taylor_QN_var_predicted_composition_4000_full.txt', sample_size = 4000)
par_quad = tl.get_val_ind_sample_file('out_files/TL_quad_p_partition_4000.txt', sample_size = 4000)
comp_quad = tl.get_val_ind_sample_file('out_files/TL_quad_p_composition_4000.txt', sample_size = 4000)
b_obs, b_par, b_comp, b_type = [], [], [], []
p_obs, p_par, p_comp = [], [], []
pcurv_obs, pcurv_par, pcurv_comp = [], [], []
r2_obs, r2_par, r2_comp = [], [], []
for study in np.unique(var_par_1000['study']):
b_obs.append((tl_pars_par['b_obs'][tl_pars_par['study'] == study])[0])
p_obs.append((tl_pars_par['p_obs'][tl_pars_par['study'] == study])[0])
r2_obs.append((tl_pars_par['R2_obs'][tl_pars_par['study'] == study])[0])
pcurv_obs.append((par_quad['emp_val'][par_quad['study'] == study])[0])
b_type.append((study_info['type'][study_info['study'] == study])[0])
sample_par = var_par[var_par['study'] == study]
sample_comp = var_comp[var_comp['study'] == study]
for i in xrange(4000):
sample_par_i = [sample_par[x][i + 5] for x in xrange(len(sample_par))]
mean_par = [sample_par['mean'][p] for p in xrange(len(sample_par)) if sample_par_i[p] != 0]
sample_par_i = [sample_par_i[p] for p in xrange(len(sample_par_i)) if sample_par_i[p] != 0]
b_par_i, inter_par_i, r_par_i, p_par_i, std_err_par_i = stats.linregress(np.log(mean_par), np.log(sample_par_i))
b_par.append(b_par_i)
p_par.append(p_par_i)
r2_par.append(r_par_i ** 2)
pcurv_par.append(par_quad[par_quad['study'] == study][0][i + 2])
sample_comp_i = [sample_comp[y][i + 5] for y in xrange(len(sample_comp))]
mean_comp = [sample_comp['mean'][q] for q in xrange(len(sample_comp)) if sample_comp_i[q] != 0]
sample_comp_i = [sample_comp_i[q] for q in xrange(len(sample_comp_i)) if sample_comp_i[q] != 0]
b_comp_i, inter_comp_i, r_comp_i, p_comp_i, std_err_comp_i = stats.linregress(np.log(mean_comp), np.log(sample_comp_i))
b_comp.append(b_comp_i)
p_comp.append(p_comp_i)
r2_comp.append(r_comp_i ** 2)
pcurv_comp.append(comp_quad[comp_quad['study'] == study][0][i + 2])
fig = plt.figure(figsize = (7, 7))
ax_p = plt.subplot(221)
tl.plot_dens_par_comp(p_obs, p_par, p_comp, ax = ax_p, legend = True, loc = 1, vline = 0.05, xlim = [0, 0.2])
ax_p.annotate('(A)', xy = (0.05, 0.92), xycoords = 'axes fraction', fontsize = 10)
plt.xlabel('p-value for b', fontsize = 8)
plt.ylabel('Density', fontsize = 8)
ax_curv = plt.subplot(222)
pcurv_par = [x for x in pcurv_par if not np.isnan(x)]
pcurv_comp = [x for x in pcurv_comp if not np.isnan(x)]
tl.plot_dens_par_comp(pcurv_obs, pcurv_par, pcurv_comp, ax = ax_curv, vline = 0.05, xlim = [0, 1])
ax_curv.annotate('(B)', xy = (0.05, 0.92), xycoords = 'axes fraction', fontsize = 10)
plt.xlabel('p-value for quadratic term', fontsize = 8)
plt.ylabel('Density', fontsize = 8)
ax_r2 = plt.subplot(223)
tl.plot_dens_par_comp(r2_obs, r2_par, r2_comp, ax = ax_r2, xlim = [0, 1])
ax_r2.annotate('(C)', xy = (0.05, 0.92), xycoords = 'axes fraction', fontsize = 10)
plt.xlabel(r'$R^2$', fontsize = 8)
plt.ylabel('Density', fontsize = 8)
ax_b = plt.subplot(224)
tl.plot_dens_par_comp(b_obs, b_par, b_comp, ax = ax_b, xlim = [0, 4])
ax_b.annotate('(D)', xy = (0.05, 0.92), xycoords = 'axes fraction', fontsize = 10)
plt.xlabel('Exponent b', fontsize = 8)
plt.ylabel('Density', fontsize = 8)
plt.savefig('FigB2.pdf', dpi = 600)