-
Notifications
You must be signed in to change notification settings - Fork 1
/
genSynthData.py
249 lines (187 loc) · 9.84 KB
/
genSynthData.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
243
244
245
246
247
248
import numpy as np
import os
import pickle
#from GPModel import *
from env import *
import scipy
import scipy.stats
from auxFunc import *
import ParHierModel
import pandas as pd
import auxFunc
def generateDataJMD(nrSubjLong, nrBiomk, nrTimepts, shiftsLowerLim, shiftsUpperLim, model,
outFolder, fileName, forceRegenerate, localParams, scalingBiomk2B, ctlDiagNr, patDiagNr):
''' generates data from a hierarchical model of disease '''
if os.path.isfile(fileName) and not forceRegenerate:
dataStruct = pickle.load(open(fileName, 'rb'))
localParams = dataStruct['localParams']
localParams['labels'] = ['biomarker %d' % d for d in range(nrBiomk)]
else:
np.random.seed(1)
# generate subject data
subShiftsLongTrue = np.random.uniform(shiftsLowerLim, shiftsUpperLim, (nrSubjLong,))
nrSubjCross = nrTimepts * nrSubjLong
# ageAtBlScanLong = np.random.uniform(lowerAgeLim,upperAgeLim, (nrSubjLong,))
# ageAtScanCross = np.zeros(nrSubjCross, float)
# ageAtBlScanCross = np.zeros(nrSubjCross, float)
yearsSinceBlScanCross = np.zeros(nrSubjCross, float)
dataCrossSB = np.zeros((nrSubjCross, nrBiomk), float)
subShiftsCrossTrue = np.zeros(nrSubjCross, float)
partCodeCross = np.zeros(nrSubjCross, float)
partCodeLong = np.array(range(nrSubjLong)) # unique id for every participant
scanTimeptsCross = np.zeros(nrSubjCross, float)
counter = 0
long2crossInd = np.zeros(nrSubjCross, int)
cross2longInd = [0 for s in range(nrSubjLong)]
for s in range(nrSubjLong):
cross2longInd[s] = np.array(range(counter,counter+nrTimepts))
for tp in range(nrTimepts):
# get currTimept, age at curr Timepints, and partCodeCross
partCodeCross[counter] = partCodeLong[s]
scanTimeptsCross[counter] = tp
yearsSinceBlScanCross[counter] = tp
# ageAtScanCross[counter] = ageAtBlScanLong[s] + tp # add one year at each timepoint
# ageAtBlScanCross[counter] = ageAtBlScanLong[s] # keep baseline age even for followups
subShiftsCrossTrue[counter] = subShiftsLongTrue[s]
long2crossInd[counter] = s
counter += 1
# yearsSinceBlScan = ageAtScanCross - ageAtBlScanCross
# generate data - find dps from age
dpsCross = yearsSinceBlScanCross + subShiftsCrossTrue # disease progression score
# dpsLongSV = [dpsCross[cross2longInd[s]] for s in range(nrSubjLong)]
diagCross = generateDiag(dpsCross, ctlDiagNr=ctlDiagNr, patDiagNr=patDiagNr)
print('diagCross', diagCross)
assert np.unique(diagCross).shape[0] >= 2
# print(adsa)
print('subShiftsCrossTrue', subShiftsCrossTrue)
# make the shifts identifiable. set origin dps=0 as the line that best separates CTL vs AD
subShiftsCrossTrue, shiftTransform = makeShiftsIdentif(
subShiftsCrossTrue, yearsSinceBlScanCross, diagCross, ctlDiagNr=ctlDiagNr, patDiagNr=patDiagNr)
print('subShiftsCrossTrue', subShiftsCrossTrue)
# print(asa)
subShiftsLongTrue = makeLongFromCross(subShiftsCrossTrue, cross2longInd)
yearsSinceBlScanLong = makeLongFromCross(yearsSinceBlScanCross, cross2longInd)
dpsCross = yearsSinceBlScanCross + subShiftsCrossTrue
dpsLongSV = makeLongFromCross(dpsCross, cross2longInd)
# print('dpsCross', dpsCross)
# print('dpsRange', np.min(dpsCross), np.max(dpsCross))
# print('subShiftsLongTrue', subShiftsLongTrue)
# print(asds)
# trueParams = dict(subShiftsLongTrue=subShiftsLongTrue,
# subShiftsCrossTrue=subShiftsCrossTrue, dpsLongSV=dpsLongSV, dpsCross=dpsCross)
#### now generate the actual biomarker data #######
dataCrossSB = model.genDataIID(dpsCross)
dataCrossSB = auxFunc.applyScalingToBiomk(dataCrossSB, scalingBiomk2B)
assert (not np.any(np.isnan(dataCrossSB)))
labels = ['biomarker %d' % d for d in range(nrBiomk)]
longData, longDiagAllTmpts, longDiag, longScanTimepts, longPartCode, longAgeAtScan, \
inverseMap, filtData, filtDiag, filtScanTimetps, filtPartCode, filtYearsSinceBlScanCross, \
= createLongData(dataCrossSB, diagCross, scanTimeptsCross,
partCodeCross, yearsSinceBlScanCross)
# localParams['data'] = dataCrossSB
# localParams['scanTimepts'] = scanTimeptsCross
# localParams['partCode'] = partCodeCross
# localParams['ageAtScan'] = ageAtScanCross
# localParams['ageAtBlScan'] = ageAtBlScanCross
# localParams['longData'] = longData
# localParams['longDiag'] = longDiag
# localParams['longScanTimepts'] = longScanTimepts
# localParams['longPartCode'] = longPartCode
# localParams['longAgeAtScan'] = longAgeAtScan
# localParams['inverseMap'] = inverseMap
# monthsSinceBlScan = 12*(ageAtScanCross - ageAtBlScanCross)
# put everything in Marco's format.
# X - list of length NR_BIOMK. X[b] - list of NR_SUBJ_LONG X[b][s] - list of visit months for subject s and biomarker b
# Y - list of length NR_BIOMK. Y[b] - list of NR_SUBJ_LONG Y[b][s] - list of biomarker values for subject s and biomarker b
# RID - list of length NR_SUBJ_LONG
X, Y, RID, visitIndices = convertToMarcoFormat(dataCrossSB, labels, yearsSinceBlScanCross, partCodeCross, diagCross)
# print(len(X[0]), len(Y[0]), RID.shape, len(visitIndices[0]))
# print(adsa)
localParams['X'] = X
localParams['Y'] = Y
localParams['RID'] = RID
localParams['labels'] = labels
localParams['visitIndices'] = visitIndices
print('RID', RID)
print('X[0][1]', X[0][1])
diagMarcoFormat = np.zeros(RID.shape)
subShiftsTrueMarcoFormatS = np.zeros(RID.shape)
subShiftsLongTrue1D = np.array([x[0] for x in subShiftsLongTrue])
for r in range(len(RID)):
print(partCodeCross == RID[r])
print(diagCross[partCodeCross == RID[r]])
diagMarcoFormat[r] = diagCross[partCodeCross == RID[r]][0]
subShiftsTrueMarcoFormatS[r] = subShiftsLongTrue1D[partCodeLong == RID[r]]
localParams['diag'] = diagMarcoFormat
biomkInFuncUnit = localParams['biomkInFuncUnit']
nrFuncUnits = localParams['nrFuncUnits']
# disease agnostic
trueDysfuncXsX = np.linspace(0, 1, num=50)
trueTrajFromDysXB = model.predPopFromDysfunc(trueDysfuncXsX)
trueTrajFromDysXB = auxFunc.applyScalingToBiomk(trueTrajFromDysXB, scalingBiomk2B)
trueSubjDysfuncScoresSU = model.predPopDys(subShiftsTrueMarcoFormatS) # dysf scores over DPS
trueParamsFuncUnits = [0 for _ in range(nrFuncUnits)]
for f in range(nrFuncUnits):
trueParamsFuncUnits[f] = dict(xsX=trueDysfuncXsX, ysXB=trueTrajFromDysXB[:, biomkInFuncUnit[f]],
subShiftsS=trueSubjDysfuncScoresSU[:,f], scalingBiomk2B=scalingBiomk2B[:, biomkInFuncUnit[f]])
# disease specific
trueLineSpacedDPSsX = np.linspace(np.min(dpsCross), np.max(dpsCross), num=50) # DPS in disease space
trueTrajPredXB = model.predPop(trueLineSpacedDPSsX) # biomk trajectory over DPS
trueTrajPredXB = auxFunc.applyScalingToBiomk(trueTrajPredXB, scalingBiomk2B)
trueDysTrajFromDpsXU = model.predPopDys(trueLineSpacedDPSsX) # dysf traj over DPS
trueParamsDis = dict(xsX=trueLineSpacedDPSsX, ysXU=trueDysTrajFromDpsXU, ysXB=trueTrajPredXB,
subShiftsS=subShiftsTrueMarcoFormatS, scalingBiomk2B=scalingBiomk2B)
localParams['trueParamsFuncUnits'] = trueParamsFuncUnits
localParams['trueParamsDis'] = trueParamsDis # add more diseases later
os.system('mkdir -p %s' % outFolder)
outFileFull = '%s/%s' % (outFolder, fileName)
pickle.dump(localParams, open(outFileFull, 'wb'), protocol=pickle.HIGHEST_PROTOCOL)
return localParams
def convertToMarcoFormat(data, labels, yearsSinceBlScan, partCode, diag):
df = pd.DataFrame(data,columns = labels)
df.insert(0, 'diag', diag)
df.insert(0, 'Month_bl', yearsSinceBlScan)
df.insert(0, 'RID', partCode)
df.insert(0, 'SUB', np.array(range(data.shape[0])))
X,Y,RID,list_biomarkers, diag, visitIndices = auxFunc.convert_table_marco(df, list_biomarkers=labels)
return X,Y,np.array(RID), visitIndices
def generateDiag(dpsCross, ctlDiagNr, patDiagNr, diagPrecDef = 0.4, muScale = 1):
nrSubjCross = dpsCross.shape[0]
controlDiagPrec = diagPrecDef
patientDiagPrec = diagPrecDef
minDps = np.min(dpsCross)
maxDps = np.max(dpsCross)
#dpsUpperLim = upperAgeLim # after this dps limit limit almost all of diags will be patient
# precision values they cannot be 1(perfect recision) as the exponential distribution is not well - defined anymore
assert (controlDiagPrec != 1 and patientDiagPrec != 1)
# multiplying the mean with nrTimepts scales perfectly to more biomk, tested on 18 / 03 / 2016
muExpoCTL = minDps + muScale * (maxDps - minDps) * (1 - controlDiagPrec**(1 / 2))
muExpoPAT = minDps + muScale * (maxDps - minDps) * (1 - patientDiagPrec**(1 / 2))
diagCross = ctlDiagNr * np.ones(nrSubjCross, int)
probControl = np.zeros(nrSubjCross, float)
for s in range(nrSubjCross):
# generate diag
dpsCurr = dpsCross[s]
probControl[s] = calcProbControlFromExpo(dpsCurr, muExpoCTL, muExpoPAT, minDps, maxDps)
if np.random.rand(1, 1) > probControl[s]:
diagCross[s] = patDiagNr
# plot probControl over dps's
nrStages = 100
stageRange = np.linspace(minDps, maxDps, nrStages)
probControlStages = np.zeros(nrStages, float)
for st in range(nrStages):
probControlStages[st] = calcProbControlFromExpo(stageRange[st], muExpoCTL, muExpoPAT, minDps, maxDps)
assert not np.isnan(probControl).any()
# print('dpsCross', dpsCross)
# print('probControl', probControl)
# print(stageRange, probControlStages)
# print(muExpoCTL, muExpoPAT)
# print(minDps, maxDps)
# pl.plot(stageRange, probControlStages)
# pl.show()
return diagCross
def calcProbControlFromExpo(stage, muExpoCTL, muExpoPAT, stageLowerLim, stageUpperLim):
probControl = scipy.stats.expon.pdf(stage-stageLowerLim, scale=muExpoCTL-stageLowerLim) / \
(scipy.stats.expon.pdf(stage-stageLowerLim, scale=muExpoCTL-stageLowerLim) +
scipy.stats.expon.pdf(stageUpperLim - stage, scale=muExpoPAT-stageLowerLim))
return probControl