def __init__(self, desk):
self.desk = desk
self.dr = desk.dr
''' ------------------------------------------------------- '''
self.entFile = bioClass.Entropy_File()
self.gg = graphPac.GeneralGraphic(desk)
self.minmax = desk.minmax
'''
Created on Sep 14, 2015
@author: Flavio Lichtenstein
@local: Unifesp DIS - Bioinformatica
for each species calc if it is p-value < .05 on its JSD distribution
'''
import classes.BarGraphic as graphPac
gg = graphPac.GeneralGraphic()
import tests.Desktop_test as Desktop_test
import random
import classes.Drosophila as dro
organism = "Drosophila"
gene = "Adh"
title = ''
cutoffLength = 100
cutoffNumSeq = 10
filename_default = 'default.ini'
if organism == "Drosophila":
dr = dro.Drosophila()
else:
dr = None
desk = Desktop_test.Desktop(filename_default, organism, gene, title,
cutoffLength, cutoffNumSeq)
def save_tables_show_graph(self,cmp_file=""):
bias_corr_list = [self.withCorrection]
print("-----------------------------------------------------------")
for bias_corr in bias_corr_list:
iLoop = 0
speciesParams = []
listMI_Anova = []
for mat in self.list_species_name_hmi:
iLoop += 1
filename, name, species, hmi, sufix = mat
numOfSeqs = self.sampleElems
L = self.LH0
print("%i) %s - %s, %s"%(iLoop, name, self.minmax, "bias corr." if bias_corr else 'no corr.' ))
sp = self.which_sp(self.organism, species)
self.withCorrection = bias_corr
if self.withCorrection:
str_correction = ' (bias corr.)'
filename_correction = '_bias_corr'
else:
str_correction = ''
filename_correction = ''
if self.frame == 0:
sFrame = ''
else:
sFrame = ', frame=' + str(self.frame)
title = 'HMI: %s %s, %s %s\n%s%s; %i seqs; len=%i%s; #letter=%i'%\
(self.organism, species, self.seqType, self.gene, self.minmax, str_correction, numOfSeqs, L, sFrame, self.numOfLetters)
''' HMI mean and sdv for each k from n sequences '''
if not self.withCorrection:
arrayMIthis = np.array(hmi.arrayMI)
arraySE = np.array(hmi.arraySE)
else:
arrayMIthis = np.array(hmi.arrayMIcorr)
arraySE = np.array(hmi.arraySEcorr)
if self.mnat:
arrayMIthis *= 1000
arraySE *= 1000
roundVal = 1
self.unit = 'mnat'
else:
roundVal = 4
self.unit = 'nat'
if self.norm:
arrayMIthis /= self.numOfLetters
arraySE /= self.numOfLetters
# self.showmsg_obs(' >>> species %s with %i sequences'%(species, numOfSeqs))
''' first time build abcissa
if choose == 1 and self.frame == 1
or self.frame == 0
'''
xSeq = []
if self.frame < 2:
for k in range( len(arrayMIthis)):
xSeq.append(k+3)
else:
''' frame1: k = {3,6,9, ... 3n} '''
for k in range( len(arrayMIthis)):
xSeq.append((k+1)*3)
arrayVal = []
ySeqSEFinal = []
''' if only one curve '''
if self.frame < 2:
arrayVal = arrayMIthis
ySeqSEFinal = arraySE
else:
''' if choose > 1 there are 3 others curves, and k jumps 3 x 3 '''
for m in range(len(arrayMIthis)):
arrayVal.append(arrayMIthis[m])
arrayVal.append(arrayMIthis[m])
arrayVal.append(arrayMIthis[m])
ySeqSEFinal.append(arraySE[m])
ySeqSEFinal.append(arraySE[m])
ySeqSEFinal.append(arraySE[m])
self.meanY = np.round(np.mean(arrayVal), roundVal)
self.medianY = np.round(np.median(arrayVal), roundVal)
self.stdY = np.round(np.std(arrayVal), roundVal)
self.maxY = np.round(np.max(arrayVal), roundVal)
self.minY = np.round(np.min(arrayVal), roundVal)
speciesParams.append([species,sp,numOfSeqs, self.meanY, self.medianY, self.stdY, self.maxY, self.minY])
listMI_Anova.append(arrayVal)
if self.showGraph or self.saveGraph:
if self.frame < 2:
gHist = graphPac.Histogram_FreqDistribution(self, title)
gHist.meanY = self.meanY
gHist.medianY = self.medianY
gHist.stdY = self.stdY
gHist.maxY = self.maxY
gHist.minY = self.minY
gHist.desk = self
gHist.plot_H_MI('V', xSeq=xSeq, ySeq=arrayVal, ySE=arraySE, showError=True, \
unit=self.unit, roundVal=roundVal)
gHist.plot_H_MI('H', xSeq=xSeq, ySeq=arrayVal, ySE=ySeqSEFinal, showError=True, \
unit=self.unit, roundVal=roundVal)
gHist.densityBar("H", seq=arrayVal, unit=self.unit, roundVal = roundVal)
else:
gHist.sameBar(xSeq=xSeq, arrayMIthis=arrayVal,linestyleCode=self.arrLinestyleCode[self.frame], color=self.arrColor[self.frame])
if (self.frame == 0) or (self.frame==3):
''' sample in filename '''
pictureName = 'HMI_%s%s%s'%(sufix, filename_correction, cmp_file)
gHist.myPlot.print_graph(self, gHist.fig, pictureName, frame=self.tk_root, stay=True)
'''
del gHist
plt.cla()
plt.clf()
gc.collect()
'''
if self.saveData :
sufix = ('%s_%s_%s_%s_frame%i_NOL%i_%iL_cutoff%i') %\
(self.organism, self.minmax, self.seqType, self.gene, self.frame, self.numOfLetters, self.cutoffLength, self.cutoffNumSeq)
''' summary with sample in filename '''
filename = 'HMI_summary_%s%s%s.txt' % (sufix, filename_correction, cmp_file)
stri = hmi.ent.calc_anova(listMI_Anova, sufix)
hmi.ent.print_data_summary(self, speciesParams, roundVal=roundVal, filename=filename, stri=stri)
self.failure = False
self.error_msg = 'Task ended. All right.'
''' in the end of save_tables_show_graph clean memory '''
del self.list_species_name_hmi
try:
gc.collect()
except:
pass
return True
def save_tables_show_graph(self, desk):
if desk.each_all == 'each':
bias_corr_list = [desk.withCorrection]
else:
bias_corr_list = [False, True]
print("-----------------------------------------------------------")
for bias_corr in bias_corr_list:
iLoop = 0
speciesParams = []
listMI_Anova = []
for mat in self.list_species_name_hmi:
iLoop += 1
filename, name, species, hmi, sufix = mat
mat = desk.dicParams[species]
if desk.minmax == 'mincut':
numOfSeqs = mat[2]
L = mat[5]
else:
numOfSeqs = mat[3]
L = mat[6]
print("%i) %s - %s, %s" %
(iLoop, name, desk.minmax,
"bias corr." if bias_corr else 'no corr.'))
if desk.organism == 'Drosophila':
sp = self.dr.mnemonic(species.replace('Drosophila ', ''))
else:
sp = species
desk.withCorrection = bias_corr
if desk.withCorrection:
str_correction = ' (bias corr.)'
filename_correction = '_bias_corr'
else:
str_correction = ''
filename_correction = ''
if desk.frame == 0:
sFrame = ''
else:
sFrame = ', frame=' + str(desk.frame)
std_rand = "" if (desk.label_random
== "") else " " + desk.label_random
title = 'HMI: %s %s, %s %s\n%s%s%s; %i seqs; len=%i%s; #letter=%i'%\
(desk.organism, species, desk.seqType, desk.gene_title, desk.minmax, str_correction, std_rand, numOfSeqs, L, sFrame, desk.numOfLetters)
''' HMI mean and sdv for each k from n sequences '''
if not desk.withCorrection:
arrayMIthis = np.array(hmi.arrayMI)
arraySE = np.array(hmi.arraySE)
else:
arrayMIthis = np.array(hmi.arrayMIcorr)
arraySE = np.array(hmi.arraySEcorr)
if desk.mnat:
arrayMIthis *= 1000
arraySE *= 1000
roundVal = 1
unit = 'mnat'
else:
roundVal = 4
unit = 'nat'
if desk.norm:
arrayMIthis /= desk.numOfLetters
arraySE /= desk.numOfLetters
# desk.showmsg_obs(' >>> species %s with %i sequences'%(species, numOfSeqs))
''' first time build abcissa
if choose == 1 and desk.frame == 1
or desk.frame == 0
'''
xSeq = []
if desk.frame < 2:
for k in range(len(arrayMIthis)):
xSeq.append(k + 3)
else:
''' frame1: k = {3,6,9, ... 3n} '''
for k in range(len(arrayMIthis)):
xSeq.append((k + 1) * 3)
arrayVal = []
ySeqSEFinal = []
''' if only one curve '''
if desk.frame < 2:
arrayVal = arrayMIthis
ySeqSEFinal = arraySE
else:
''' if choose > 1 there are 3 others curves, and k jumps 3 x 3 '''
for m in range(len(arrayMIthis)):
arrayVal.append(arrayMIthis[m])
arrayVal.append(arrayMIthis[m])
arrayVal.append(arrayMIthis[m])
ySeqSEFinal.append(arraySE[m])
ySeqSEFinal.append(arraySE[m])
ySeqSEFinal.append(arraySE[m])
self.meanY = np.round(np.mean(arrayVal), roundVal)
self.medianY = np.round(np.median(arrayVal), roundVal)
self.stdY = np.round(np.std(arrayVal), roundVal)
self.maxY = np.round(np.max(arrayVal), roundVal)
self.minY = np.round(np.min(arrayVal), roundVal)
speciesParams.append([
species, sp, numOfSeqs, self.meanY, self.medianY,
self.stdY, self.maxY, self.minY
])
listMI_Anova.append(arrayVal)
if desk.showGraph or desk.saveGraph:
if desk.frame < 2:
gHist = graphPac.Histogram_FreqDistribution(
desk, title)
gHist.meanY = self.meanY
gHist.medianY = self.medianY
gHist.stdY = self.stdY
gHist.maxY = self.maxY
gHist.minY = self.minY
gHist.plot_H_MI('V',
xSeq=xSeq,
ySeq=arrayVal,
ySE=arraySE,
showError=True)
gHist.plot_H_MI('H',
xSeq=xSeq,
ySeq=arrayVal,
ySE=ySeqSEFinal,
showError=True)
gHist.densityBar("H", seq=arrayVal)
else:
gHist.sameBar(
xSeq=xSeq,
arrayMIthis=arrayVal,
linestyleCode=self.arrLinestyleCode[desk.frame],
color=self.arrColor[desk.frame])
if (desk.frame == 0) or (desk.frame == 3):
pictureName = 'HMI_%s%s' % (sufix, filename_correction)
gHist.myPlot.print_graph(desk,
gHist.fig,
pictureName,
frame=desk.tk_root)
'''
del gHist
plt.cla()
plt.clf()
gc.collect()
'''
if desk.saveData:
sufix = ('%s_%s_%s_%s_frame%i_NOL%i_%iL_cutoff%i') %\
(desk.organism, desk.minmax, desk.seqType, desk.gene_title, desk.frame, desk.numOfLetters, desk.cutoffLength, desk.cutoffNumSeq)
std_rand = "" if (desk.label_random
== "") else desk.stri_random
filename = 'HMI_summary_%s%s%s.txt' % (
sufix, filename_correction, std_rand)
stri = hmi.ent.calc_anova(listMI_Anova, sufix)
hmi.ent.print_data_summary(desk,
speciesParams,
roundVal=roundVal,
filename=filename,
stri=stri)
self.failure = False
self.error_msg = 'Task ended. All right.'
return True
def save_tables_show_graph(self, desk, which_module):
if desk.each_all == 'each':
list_modules = [which_module]
bias_corr_list = [desk.withCorrection]
else:
list_modules = ['Entropy', 'VMI']
bias_corr_list = [False, True]
print("-----------------------------------------------------------")
for which_module in list_modules:
for bias_corr in bias_corr_list:
iLoop = 0
listMI_Anova = []
speciesParams = []
for mat in self.list_species_name_vmi:
iLoop += 1
filename, name, species, vmi, sufix = mat
mat = desk.dicParams[species]
if desk.minmax == 'mincut':
numOfSeqs = mat[2]
L = mat[5]
else:
numOfSeqs = mat[3]
L = mat[6]
print("%i) %s - %s %s %s"%(iLoop, name, desk.minmax, "bias corr." if bias_corr else 'no corr.', which_module ))
if desk.organism == 'Drosophila':
sp = self.dr.mnemonic(species.replace('Drosophila ',''))
else:
sp = species
desk.withCorrection = bias_corr
if desk.withCorrection:
str_correction = ' (bias corr.)'
filename_correction = '_bias_corr'
else:
str_correction = ''
filename_correction = ''
std_rand = "" if (desk.label_random == "") else " " + desk.label_random
if which_module == 'Entropy':
title = 'Entropy Distribution %s %s, %s %s\n%s%s%s; %i seqs; len=%i; #letter=%i'%\
(desk.organism, species, desk.seqType, desk.gene_title, desk.minmax, str_correction, std_rand, numOfSeqs, L, desk.numOfLetters)
else:
title = 'VMI Heat Map %s %s, %s %s %s%s%s\n%i seqs; len=%i; #letter=%i'%\
(desk.organism, species, desk.seqType, desk.gene_title, desk.minmax, str_correction, std_rand, numOfSeqs, L, desk.numOfLetters)
if which_module == 'Entropy':
if not desk.withCorrection:
arrayVal = np.array(vmi.HShannonList)
arraySE = np.array(vmi.SeHShannonList)
else:
arrayVal = np.array(vmi.HShannonCorrList)
arraySE = np.array(vmi.SeHShannonCorrList)
else:
if not desk.withCorrection:
arrayVal = np.array(vmi.MIlist)
arraySE = np.array(vmi.SeMIList)
else:
arrayVal = np.array(vmi.MIcorrList)
arraySE = np.array(vmi.SeMICorrList)
''' normalization dividing by numOfLetters '''
if desk.norm:
arrayVal /= desk.numOfLetters
arraySE /= desk.numOfLetters
''' mili nats '''
if desk.mnat:
arrayVal *= 1000
arraySE *= 1000
roundVal = 2
else:
roundVal = 4
# print '--- params -----------------'
maxMI = 0
maxiPos = None
is_zero = True
for pos in range(len(arrayVal)):
if arrayVal[pos] > maxMI:
maxMI = arrayVal[pos]
SE = arraySE[pos]
i,j = vmi.ijList[pos]
maxiPos = [i,j,maxMI,SE]
is_zero = False
if is_zero:
stri = '### Species %s has MI = ZERO. Too conserved data sequences. Impossible to include in analysis.'%(species)
print(stri)
continue
xSeq = [x for x in range(len(arrayVal))]
self.meanY = np.round(np.mean(arrayVal), roundVal)
self.medianY = np.round(np.median(arrayVal), roundVal)
self.stdY = np.round(np.std(arrayVal), roundVal)
self.maxY = np.round(np.max(arrayVal), roundVal)
self.minY = np.round(np.min(arrayVal), roundVal)
speciesParams.append([species,sp,numOfSeqs, self.meanY, self.medianY, self.stdY, self.maxY, self.minY])
if desk.saveData:
listMI_Anova.append(arrayVal)
if desk.showGraph or desk.saveGraph:
if which_module == 'Entropy':
gHist = graphPac.Histogram_FreqDistribution(desk, title)
gHist.meanY = self.meanY
gHist.medianY = self.medianY
gHist.stdY = self.stdY
gHist.maxY = self.maxY
gHist.minY = self.minY
gHist.plot_H_MI('V', xSeq=xSeq, ySeq=arrayVal, ySE=arraySE, showError=True)
gHist.densityBar('V', seq=arrayVal)
pictureName = 'VHShannon_%s%s%s'%(sufix, filename_correction, desk.stri_random)
gHist.myPlot.print_graph(desk, gHist.fig, pictureName, frame=desk.tk_root)
else:
''' 3D dont has ceil '''
if desk.is3D:
limSup = self.maxY
else:
ceil = desk.heatmap_ceil_value
''' the same roof for all heatmaps: parametrize in future '''
if desk.heatmap_ceil:
if self.maxY <= ceil:
limSup = ceil
else:
limSup = self.maxY
else:
limSup = self.maxY
'''' updated 28/09/2015 '''
gMI = vmi.plotHeatMap(desk, desk.is3D, arrayVal, vmi.ijList, L, maxiPos, title, species=species, limSup=limSup, roundVal=roundVal, str_correction=str_correction)
if desk.is3D:
pictureName = 'HeatMap_3D_VMI_%s%s%s'%(sufix, filename_correction,desk.stri_random)
else:
gMI.densityHeatmapBar(arrayVal, limSup)
pictureName = 'HeatMap_2D_VMI_%s%s%s'%(sufix, filename_correction,desk.stri_random)
gMI.myPlot.print_graph(desk, gMI.fig, pictureName, frame=desk.tk_root)
'''
plt.cla()
plt.clf()
plt.close()
del gMI
gc.collect()
'''
# Flavio 02/06/2015
if which_module == 'Entropy':
which_symb = 'VHShannon'
else:
which_symb = 'VMI'
if desk.saveData:
sufix = ('%s_%s_%s_%s_NOL%i_%iL_cutoff%i') %\
(desk.organism, desk.minmax, desk.seqType, desk.gene_title, desk.numOfLetters, desk.cutoffLength, desk.cutoffNumSeq)
filename = '%s_summary_%s%s%s.txt' % (which_symb, sufix, filename_correction, desk.stri_random)
stri = vmi.ent.calc_anova(listMI_Anova, sufix)
# desk, speciesParams, roundVal=4, filename=None, stri = '', saveData=False):
vmi.ent.print_data_summary(desk, speciesParams, roundVal=roundVal, filename=filename, stri=stri)
def looping(self, desk, opt):
plt.close("all")
plt.clf()
desk.withCorrection = opt[0]
desk.minmax = opt[1]
if desk.withCorrection:
self.str_correction = '-bias corr.'
self.filename_correction = '_bias_corr'
else:
self.str_correction = ''
self.filename_correction = ''
print("\n--->>>", desk.minmax, self.str_correction)
desk.colorThreshold = desk.colorThreshold_var.get()
if desk.mnat:
desk.unit = 'mnat'
desk.factor = 1000
desk.roundVal = 2
else:
desk.unit = 'nat'
desk.factor = 1
desk.roundVal = 4
if desk.vert_horiz == 'HMI':
xLabel = 'JSD(HMI) (%s)' % (desk.unit)
title = "Hierarchical Cluster Method=%s of JSD(HMI)- %s %s %s"\
%(desk.cluster_method_desc, desk.organism, desk.seqType, desk.gene)
if desk.frame > 0:
title += '\nJSD(HMI) %s%s, desk.frame %i, #letter %i, min(L)=%i, min(#seqs)=%i' % \
(desk.minmax, self.str_correction, desk.frame, desk.numOfLetters, desk.cutoffLength, desk.cutoffNumSeq)
else:
title += '\n%s%s, letter %i, min(L)=%i, min(#seqs)=%i' % \
(desk.minmax, self.str_correction, desk.numOfLetters, desk.cutoffLength, desk.cutoffNumSeq)
elif desk.vert_horiz == 'VMI':
xLabel = 'JSD(VMI) (%s)' % (desk.unit)
''' multidimensional distance '''
title = "Hierarchical Cluster Method=%s of JSD(VMI), %s %s %s"\
%(desk.cluster_method_desc, desk.organism, desk.seqType, desk.gene)
title += '\n%s%s, #letter %i, min(L)=%i, min(#seqs)=%i' % \
(desk.minmax, self.str_correction, desk.numOfLetters, desk.cutoffLength, desk.cutoffNumSeq)
else:
xLabel = 'JSD(VSH) (nat)'
''' multidimensional distance '''
title = "Hierarchical Cluster Method=%s of JSD(VSH), %s %s %s"\
%(desk.cluster_method_desc, desk.organism, desk.seqType, desk.gene)
title += '\n%s%s, #letter %i, min(L)=%i, min(#seqs)=%i' % \
(desk.minmax, self.str_correction, desk.numOfLetters, desk.cutoffLength, desk.cutoffNumSeq)
desk.set_cluster_filenames()
filename = desk.cluster_input_filename
ret, _, colHeaders, dataMatrix = self.open_distance_matrix_file(
desk.rootTable + filename)
if not ret:
self.error_msg = 'Could not find %s' % (desk.rootTable + filename)
return False
pictureName = 'Cluster_' + filename.replace('.txt', '')
''' desk.dr defined in pipe_desktop get_params() '''
if desk.dr:
rows = desk.dr.labels(colHeaders)
else:
rows = colHeaders
#convert native python array into a numpy array
# dataMatrix = log10(dataMatrix)
# print dataMatrix
dataMatrix = np.array(dataMatrix)
maxDist = 0
if desk.factor != 1:
for i in range(len(dataMatrix)):
for j in range(len(dataMatrix[i])):
dataMatrix[i][j] = dataMatrix[i][j] * desk.factor
if dataMatrix[i][j] > maxDist:
maxDist = dataMatrix[i][j]
else:
for i in range(len(dataMatrix)):
for j in range(len(dataMatrix[i])):
if dataMatrix[i][j] > maxDist:
maxDist = dataMatrix[i][j]
# single, weighted, average, co mplete
linkageMatrix = linkage(dataMatrix,
method=desk.cluster_method,
metric='euclidean')
''' finding maximum '''
maxLinkDist = 0
for i in range(len(linkageMatrix)):
for j in range(len(linkageMatrix[i])):
if linkageMatrix[i][j] > maxLinkDist:
maxLinkDist = linkageMatrix[i][j]
''' hierarchical cluster distorce distances
factor = maxDist/(2*maxLinkDist) '''
for i in range(len(linkageMatrix)):
linkageMatrix[i][2] = round(linkageMatrix[i][2] * .5,
desk.roundVal)
fig = plt.figure(1, dpi=desk.dpi)
ax = fig.add_subplot('111')
plt.subplots_adjust(bottom=.1, left=.05, right=.84)
yLabel = 'species'
plt.rcParams['lines.linewidth'] = 2.5
fontsize = 26
plt.title(title, fontsize=fontsize)
ax.set_xlabel(xLabel, fontsize=fontsize)
ax.set_ylabel(yLabel, fontsize=fontsize)
# make colorbar labels bigger
leaf_font_size = 28
''' ddata = '''
try:
dendrogram(
linkageMatrix,
color_threshold=desk.colorThreshold,
labels=rows,
orientation='right'
) # show_leaf_counts=True , leaf_font_size=leaf_font_size
except:
print("Failed in printing dendrogram")
pass
plt.xticks(fontsize=leaf_font_size)
plt.yticks(fontsize=leaf_font_size)
'''
# print ddata
spList = ddata['ivl']
# print len(spList), spList
nickList = copy.deepcopy(spList)
nickList.sort()
dic = {}
for i in range(len(spList)):
sp = spList[i]
for j in range(len(nickList)):
if sp == nickList[j]:
dic[i] = j
#print i, spList[i], ' equal ',j, nickList[j]
break
count = 0
for i, d in zip(ddata['icoord'], ddata['dcoord']):
count += 1
# print i, d
# specie01 x specie02 - mean error distance
num = (i[0]-5)/10.
sp1a = int(num)
diff = num - sp1a
if diff == 0:
wei1a = 1
sp1b = sp1a
wei1b = 0
else:
sp1b = sp1a+1
wei1a = diff
wei1b = 1. - wei1a
#if num == 0:
# print '>>>> viri'
num = (i[2]-5)/10.
sp2a = int(num)
diff = num - sp2a
if diff == 0:
sp2b = sp2a
wei2a = 1
wei2b = 0
else:
sp2b = sp2a+1
wei2a = diff
wei2b = 1. - wei2a
#print sp1a, sp1b, sp2a, sp2b
#print wei1a, wei1b, wei2a, wei2b
ste = 0.
if wei1a>0 and wei2a>0:
ste += wei1a*wei2a*seMatrix[dic[sp1a]][dic[sp2a]]
if wei1a>0 and wei2b>0:
ste += wei1a*wei2b*seMatrix[dic[sp1a]][dic[sp2b]]
if wei1b>0 and wei2a>0:
ste += wei1b*wei2a*seMatrix[dic[sp1b]][dic[sp2a]]
if wei1b>0 and wei2b>0:
# print sp1b, sp2b
ste += wei1b*wei2b*seMatrix[dic[sp1b]][dic[sp2b]]
ste = round(ste,4)
dist = seMatrix[dic[sp1a]][dic[sp2a]]
dist = round(dist,4)
# print 'dist', dist, 'ste', ste
x = 0.5 * sum(i[1:3])
y = round(d[1],4)
stry = str(y) + '\nd='+str(dist) + '\nse='+str(ste)
plt.plot(x, y, 'ro')
stry = ''
if abs(y) > desk.colorThreshold:
plt.annotate(stry, (x, y), xytext=(0, -8),
textcoords='offset points',
va='top', ha='center')
'''
self.myPlot = graphPac.Plot()
self.myPlot.print_graph(desk,
fig,
pictureName=pictureName,
frame=desk.tk_root,
stay=True)
return True
def __init__(self, desk):
self.desk = desk
self.failure = True
self.error_msg = ''
try:
desk.get_params()
except:
self.error_msg = 'Could not get parameters.'
return
mbs = mb.mrBayesClass(desk)
if not mbs.read_runPs():
self.error_msg = 'Problems reading Mr.Bayes files'
return
if not mbs.stat_files_to_dic():
self.error_msg = 'Could not read Mr.Bayes statistic files'
return
mbs.summary = {}
params = []
if desk.piA_var.get(): params.append("pi(A)")
if desk.piC_var.get(): params.append("pi(C)")
if desk.piG_var.get(): params.append("pi(G)")
if desk.piT_var.get(): params.append("pi(T)")
if desk.rAC_var.get(): params.append("r(A<->C)")
if desk.rAG_var.get(): params.append("r(A<->G)")
if desk.rAT_var.get(): params.append("r(A<->T)")
if desk.rCG_var.get(): params.append("r(C<->G)")
if desk.rCT_var.get(): params.append("r(C<->T)")
if desk.rGT_var.get(): params.append("r(G<->T")
if desk.LnL_var.get(): params.append("LnL")
if desk.LnPr_var.get(): params.append("LnPr")
if desk.TL_var.get(): params.append("TL")
if desk.alpha_var.get(): params.append("alpha")
if desk.off_on_var.get(): params.append("s(off->on)")
if desk.on_off_var.get(): params.append("s(on->off)")
if desk.pinvar_var.get(): params.append("pinvar")
if not params:
self.error_msg = 'Define at least one param.'
return
iPar = 0
numCols = 46
numLines = 16
self.myPlot = graphPac.Plot()
for par in params:
iPar += 1
mbs.summary[par] = {}
#fig = figList[iPar]
fig = plt.figure(iPar)
plt.subplots_adjust(left=0.05, right=0.95, bottom=0.05, top=0.90)
# print plt.get_backend()
mng = plt.get_current_fig_manager()
mng.window.wm_geometry("1400x900+50+50")
seqXs = []
mini, maxi = float('inf'), -float('inf')
for min_max in mbs.min_maxs:
for alig_cons in mbs.alig_conss:
for std_covar in mbs.std_covars:
study = min_max + '-' + alig_cons + '-' + std_covar
try:
vals = np.array(mbs.dic_detaild_params[study][par])
if np.min(vals) < mini:
mini = min(vals)
elif np.max(vals) > maxi:
maxi = max(vals)
except:
pass
if par != "LnL" and par != "LnPr":
if par == "alpha":
mini = 0
else:
if mini < 0: mini = 0
if par in ["pi(A)", "pi(C)", "pi(G)", "pi(T)"]:
if maxi > 1: maxi = 1
ticks = []
ticks0 = []
div = 5
delta = (maxi - mini) / float(div - 1)
for i in range(div):
ticks.append(round(mini + delta * i, 2))
ticks0.append(0)
cntCols = 0
stri_summary = " \tAligned\tConsensus\n"
for min_max in mbs.min_maxs:
for alig_cons in mbs.alig_conss:
listData = []
cntLine = 0
for std_covar in mbs.std_covars:
try:
_ = mbs.mb_filenames[min_max][alig_cons][std_covar]
didFind = True
except:
didFind = False
continue
study = min_max + '-' + alig_cons + '-' + std_covar
'''
try:
dic = mbs.dic_detaild_params[study]
except:
# if file not found or could not read: it is not in dic_detailed_params
continue
'''
mbs.summary[par][study] = {}
try:
vals = np.array(mbs.dic_detaild_params[study][par])
mbs.summary[par][study] = {}
N = len(vals)
seqXs.append(np.array(vals))
mu, sigma, hmu, hsigma, q025, _, q2, _, q975 = mbs.stat_ppf_decimal(
vals, par)
mbs.summary[par][study]["N"] = N
mbs.summary[par][study]["mu"] = mu
mbs.summary[par][study]["sigma"] = sigma
SE = sigma / np.sqrt(N)
mbs.summary[par][study]["SE"] = SE
mbs.summary[par][study]["N"] = N
mbs.summary[par][study]["median"] = q2
mbs.summary[par][study]["q025"] = q025
mbs.summary[par][study]["q975"] = q975
if hmu:
mbs.summary[par][study]["hmu"] = hmu
mbs.summary[par][study]["hsigma"] = hsigma
hSE = hsigma / np.sqrt(N)
mbs.summary[par][study]["hSE"] = hSE
mbs.summary[par][study]["min"] = np.min(vals)
mbs.summary[par][study]["max"] = np.max(vals)
try:
''' LnL doesn't have ESS and PSRF - adopted from TL
total tree length (the sum of all branch lengths, TL) '''
if par == "LnL":
mbs.summary[par][study][
"avgESS"] = mbs.dic_pstat[study]["TL"][
'avgESS']
else:
mbs.summary[par][study][
"avgESS"] = mbs.dic_pstat[study][par][
'avgESS']
except:
mbs.summary[par][study]["avgESS"] = 0
try:
if par == "LnL":
mbs.summary[par][study][
"PSRF"] = mbs.dic_pstat[study]["TL"][
'PSRF']
else:
mbs.summary[par][study][
"PSRF"] = mbs.dic_pstat[study][par][
'PSRF']
except:
mbs.summary[par][study]["PSRF"] = 10000
if mbs.summary[par][study]["avgESS"] < 90:
sAvgESS = "avgESS=%3.1f **" % mbs.summary[par][
study]["avgESS"]
else:
sAvgESS = "avgESS=%3.1f" % mbs.summary[par][
study]["avgESS"]
if (mbs.summary[par][study]["PSRF"] - 1) > .1:
sPSRF = " PSRF=%1.2f ***" % mbs.summary[par][
study]["PSRF"]
else:
sPSRF = " PSRF=%1.2f" % mbs.summary[par][
study]["PSRF"]
if cntLine == 0:
ax = plt.subplot2grid((numLines, numCols),
(0, cntCols),
rowspan=4,
colspan=10)
else:
ax = plt.subplot2grid((numLines, numCols),
(6, cntCols),
rowspan=4,
colspan=10)
if par == "LnL":
print par, study, mu, sigma, hmu, hsigma
stri_summary = mbs.show_lnl(
desk, plt, ax, study, sAvgESS, sPSRF,
cntLine, cntCols, min_max, alig_cons,
stri_summary)
else:
print par, study, mu, sigma
mbs.show_distrib(plt, ax, par, study, sAvgESS,
sPSRF, ticks, desk.colors,
iPar, mini, maxi, cntCols)
listData.append(vals)
cntLine += 1
except:
pass
if didFind:
if len(listData) == 2:
if par == "LnL":
mbs.show_LRT(alig_cons, par, plt, numLines,
numCols, cntCols)
else:
mbs.show_conf_interval(min_max, alig_cons, par,
plt, numLines, numCols,
cntCols, ticks, ticks0,
mini, maxi, listData)
mbs.show_qqplot2(min_max, alig_cons, par, plt,
numLines, numCols, cntCols,
listData)
else:
pass
cntCols += 12
if par == "LnL":
stri = ""
else:
f_value, p_value = mbs.calc_anova(seqXs)
if p_value <= 0.05:
stri = ', at least one distribution is statistically different.'
else:
stri = ', the distributions are statistically similar.'
stri += 'ANOVA: f-value %2.3e p_value %2.3e' % (f_value,
p_value)
left = .05
top = .97
fig.text(left, top, par + stri, color="red")
stri = mbs.ttest_summary(par)
print stri
sPar = par.replace(">", "").replace("<", "")
pictureName = "%s_%s_mr_bayes_analisis_param_%s" % (
desk.organism, desk.gene_title, sPar)
self.myPlot.print_graph(self.desk,
fig,
pictureName,
frame=self.desk.tk_root,
stay=False)
print stri_summary
if desk.saveData:
mbs.save_params(params)
self.failure = False
return