def getQuantAlphaData(request):
samples = Sample.objects.all()
samples.query = pickle.loads(request.session['selected_samples'])
selected = samples.values_list('sampleid')
qs1 = Sample.objects.all().filter(sampleid__in=selected)
if request.is_ajax():
allJson = request.GET["all"]
all = simplejson.loads(allJson)
button = int(all["button"])
sig_only = int(all["sig_only"])
norm = int(all["normalize"])
taxaString = all["taxa"]
taxaDict = simplejson.JSONDecoder(object_pairs_hook=multidict).decode(taxaString)
metaString = all["meta"]
metaDict = simplejson.JSONDecoder(object_pairs_hook=multidict).decode(metaString)
metaDF = quantAlphaMetaDF(qs1, metaDict)
myList = metaDF['sampleid'].tolist()
mySet = list(set(myList))
taxaDF = taxaProfileDF(mySet)
factor = 'none'
if norm == 1:
factor = 'none'
elif norm == 2:
factor = 'min'
elif norm == 3:
factor = '10th percentile'
elif norm == 4:
factor = '25th percentile'
elif norm == 5:
factor = 'median'
final_fieldList = []
for key in metaDict:
final_fieldList.append(metaDict[key])
normDF = normalizeAlpha(taxaDF, taxaDict, mySet, factor)
finalDF = metaDF.merge(normDF, on='sampleid', how='outer')
finalDF[[final_fieldList[0], 'count', 'rel_abund', 'rich', 'diversity']] = finalDF[[final_fieldList[0], 'count', 'rel_abund', 'rich', 'diversity']].astype(float)
pd.set_option('display.max_rows', finalDF.shape[0], 'display.max_columns', finalDF.shape[1], 'display.width', 1000)
finalDict = {}
seriesList = []
xAxisDict = {}
yAxisDict = {}
grouped1 = finalDF.groupby(['rank', 'taxa_name', 'taxa_id'])
for name1, group1 in grouped1:
dataList = []
x = []
y = []
if button == 1:
dataList = group1[[final_fieldList[0], 'count']].values.tolist()
x = group1[final_fieldList[0]].values.tolist()
y = group1['count'].values.tolist()
elif button == 2:
dataList = group1[[final_fieldList[0], 'rel_abund']].values.tolist()
x = group1[final_fieldList[0]].values.tolist()
y = group1['rel_abund'].values.tolist()
elif button == 3:
dataList = group1[[final_fieldList[0], 'rich']].values.tolist()
x = group1[final_fieldList[0]].values.tolist()
y = group1['rich'].values.tolist()
elif button == 4:
dataList = group1[[final_fieldList[0], 'diversity']].values.tolist()
x = group1[final_fieldList[0]].values.tolist()
y = group1['diversity'].values.tolist()
if max(x) == min(x):
stop = 0
else:
stop = 1
slope, intercept, r_value, p_value, std_err = stats.linregress(x, y)
p_value = "%0.3f" % p_value
r_square = r_value * r_value
r_square = "%0.4f" % r_square
min_y = slope*min(x) + intercept
max_y = slope*max(x) + intercept
slope = "%.3E" % slope
intercept = "%.3E" % intercept
regrList = []
regrList.append([min(x), min_y])
regrList.append([max(x), max_y])
if sig_only == 0:
seriesDict = {}
seriesDict['type'] = 'scatter'
seriesDict['name'] = name1
seriesDict['data'] = dataList
seriesList.append(seriesDict)
if stop == 0:
regDict = {}
elif stop == 1:
regrDict = {}
regrDict['type'] = 'line'
name2 = list(name1)
temp = 'R2: ' + str(r_square) + '; p-value: ' + str(p_value) + '<br>' + '(y = ' + str(slope) + 'x' + ' + ' + str(intercept)
print temp
name2.append(temp)
print name2
regrDict['name'] = name2
regrDict['data'] = regrList
seriesList.append(regrDict)
if sig_only == 1:
if p_value <= 0.05:
seriesDict = {}
seriesDict['type'] = 'scatter'
name2 = list(name1)
temp = 'R2: ' + str(r_square) + '; p-value: ' + str(p_value) + '<br>' + '(y = ' + str(slope) + 'x' + ' + ' + str(intercept)
name2.append(temp)
seriesDict['name'] = name2
seriesDict['data'] = dataList
seriesList.append(seriesDict)
regrDict = {}
regrDict['type'] = 'line'
regrDict['name'] = name1
regrDict['data'] = regrList
seriesList.append(regrDict)
xTitle = {}
xTitle['text'] = final_fieldList[0]
xAxisDict['title'] = xTitle
yTitle = {}
if button == 1:
yTitle['text'] = 'Sequence Reads'
elif button == 2:
yTitle['text'] = 'Relative Abundance'
elif button == 3:
yTitle['text'] = 'Species Richness'
elif button == 4:
yTitle['text'] = 'Shannon Diversity'
yAxisDict['title'] = yTitle
finalDict['series'] = seriesList
finalDict['xAxis'] = xAxisDict
finalDict['yAxis'] = yAxisDict
if not seriesList:
finalDict['empty'] = 0
else:
finalDict['empty'] = 1
finalDF.reset_index(drop=True, inplace=True)
res_table = finalDF.to_html(classes="table display")
res_table = res_table.replace('border="1"', 'border="0"')
finalDict['res_table'] = str(res_table)
res = simplejson.dumps(finalDict)
return HttpResponse(res, content_type='application/json')
def getQuantBetaData(request):
samples = Sample.objects.all()
samples.query = pickle.loads(request.session['selected_samples'])
selected = samples.values_list('sampleid')
qs1 = Sample.objects.all().filter(sampleid__in=selected)
if request.is_ajax():
allJson = request.GET["all"]
all = simplejson.loads(allJson)
button = int(all["button"])
taxaLevel = int(all["taxa"])
distance = int(all["distance"])
norm = int(all["normalize"])
PC1 = all["PC1"]
metaString = all["meta"]
metaDict = simplejson.JSONDecoder(object_pairs_hook=multidict).decode(metaString)
metaDF = quantBetaMetaDF(qs1, metaDict)
myList = metaDF['sampleid'].tolist()
mySet = list(set(myList))
taxaDF = taxaProfileDF(mySet)
factor = 'none'
if norm == 1:
factor = 'none'
elif norm == 2:
factor = 'min'
elif norm == 3:
factor = '10th percentile'
elif norm == 4:
factor = '25th percentile'
elif norm == 5:
factor = 'median'
normDF = normalizeBeta(taxaDF, taxaLevel, mySet, factor)
finalDF = metaDF.merge(normDF, on='sampleid', how='outer')
pd.set_option('display.max_rows', finalDF.shape[0], 'display.max_columns', finalDF.shape[1], 'display.width', 1000)
fieldList = []
for key in metaDict:
fieldList.append(metaDict[key])
matrixDF = pd.DataFrame()
if button == 1:
matrixDF = finalDF.pivot(index='taxaid', columns='sampleid', values='count')
elif button == 2:
matrixDF = finalDF.pivot(index='taxaid', columns='sampleid', values='rel_abund')
elif button == 3:
matrixDF = finalDF.pivot(index='taxaid', columns='sampleid', values='rich')
elif button == 4:
matrixDF = finalDF.pivot(index='taxaid', columns='sampleid', values='diversity')
datamtx = asarray(matrixDF[mySet].T)
numrows, numcols = shape(datamtx)
dists = zeros((numrows, numrows))
if distance == 1:
dist = pdist(datamtx, 'braycurtis')
dists = squareform(dist)
elif distance == 2:
dist = pdist(datamtx, 'canberra')
dists = squareform(dist)
elif distance == 3:
dist = pdist(datamtx, 'dice')
dists = squareform(dist)
elif distance == 4:
dist = pdist(datamtx, 'euclidean')
dists = squareform(dist)
elif distance == 5:
dist = pdist(datamtx, 'jaccard')
dists = squareform(dist)
eigvals, coordinates, proportion_explained = PCoA(dists)
numaxes = len(eigvals)
axesList = []
for i in range(numaxes):
j = i + 1
axesList.append('PC' + str(j))
valsDF = pd.DataFrame(eigvals, columns=['EigenVals'], index=axesList)
propDF = pd.DataFrame(proportion_explained, columns=['Variance Explained (R2)'], index=axesList)
eigenDF = valsDF.join(propDF)
metaDF.set_index('sampleid', drop=True, inplace=True)
pcoaDF = pd.DataFrame(coordinates, columns=axesList, index=mySet)
resultDF = metaDF.join(pcoaDF)
pd.set_option('display.max_rows', resultDF.shape[0], 'display.max_columns', resultDF.shape[1], 'display.width', 1000)
finalDict = {}
seriesList = []
xAxisDict = {}
yAxisDict = {}
dataList = resultDF[[PC1, fieldList[0]]].values.tolist()
seriesDict = {}
seriesDict['type'] = 'scatter'
seriesDict['name'] = fieldList
seriesDict['data'] = dataList
seriesList.append(seriesDict)
x = resultDF[PC1].values.tolist()
y = resultDF[fieldList[0]].values.tolist()
if max(x) == min(x):
stop = 0
else:
stop = 1
slope, intercept, r_value, p_value, std_err = stats.linregress(x, y)
p_value = "%0.3f" % p_value
r_square = r_value * r_value
r_square = "%0.4f" % r_square
min_y = slope*min(x) + intercept
max_y = slope*max(x) + intercept
slope = "%.3E" % slope
intercept = "%.3E" % intercept
regrList = []
regrList.append([min(x), min_y])
regrList.append([max(x), max_y])
if stop == 0:
regDict = {}
elif stop == 1:
regrDict = {}
regrDict['type'] = 'line'
regrDict['name'] = 'R2: ' + str(r_square) + '; p-value: ' + str(p_value) + '<br>' + '(y = ' + str(slope) + 'x' + ' + ' + str(intercept) + ')'
regrDict['data'] = regrList
seriesList.append(regrDict)
xTitle = {}
xTitle['text'] = PC1
xAxisDict['title'] = xTitle
yTitle = {}
yTitle['text'] = fieldList[0]
yAxisDict['title'] = yTitle
finalDict['series'] = seriesList
finalDict['xAxis'] = xAxisDict
finalDict['yAxis'] = yAxisDict
result = ""
result = result + '===============================================\n'
if taxaLevel == 1:
result = result + 'Taxa level: Kingdom' + '\n'
elif taxaLevel == 2:
result = result + 'Taxa level: Phyla' + '\n'
elif taxaLevel == 3:
result = result + 'Taxa level: Class' + '\n'
elif taxaLevel == 4:
result = result + 'Taxa level: Order' + '\n'
elif taxaLevel == 5:
result = result + 'Taxa level: Family' + '\n'
elif taxaLevel == 6:
result = result + 'Taxa level: Genus' + '\n'
elif taxaLevel == 7:
result = result + 'Taxa level: Species' + '\n'
if button == 1:
result = result + 'Dependent Variable: Sequence Reads' + '\n'
elif button == 2:
result = result + 'Dependent Variable: Relative Abundance' + '\n'
elif button == 3:
result = result + 'Dependent Variable: Species Richness' + '\n'
elif button == 4:
result = result + 'Dependent Variable: Shannon Diversity' + '\n'
result = result + 'Independent Variable: ' + str(fieldList[0]) + '\n'
if distance == 1:
result = result + 'Distance score: Bray-Curtis' + '\n'
elif distance == 2:
result = result + 'Distance score: Canberra' + '\n'
elif distance == 3:
result = result + 'Distance score: Dice' + '\n'
elif distance == 4:
result = result + 'Distance score: Euclidean' + '\n'
elif distance == 5:
result = result + 'Distance score: Jaccard' + '\n'
result = result + '===============================================\n'
result = result + str(eigenDF) + '\n'
result = result + '===============================================\n'
result = result + '\n\n\n\n'
finalDict['text'] = result
resultDF.reset_index(drop=True, inplace=True)
res_table = resultDF.to_html(classes="table display")
res_table = res_table.replace('border="1"', 'border="0"')
finalDict['res_table'] = str(res_table)
nameList = list(metaDF['sample_name'])
distsDF = pd.DataFrame(dists, columns=nameList, index=nameList)
dist_table = distsDF.to_html(classes="table display")
dist_table = dist_table.replace('border="1"', 'border="0"')
finalDict['dist_table'] = str(dist_table)
res = simplejson.dumps(finalDict)
return HttpResponse(res, content_type='application/json')
def getCatAlphaData(request):
### get sample list from cookie
samples = Sample.objects.all()
samples.query = pickle.loads(request.session['selected_samples'])
selected = samples.values_list('sampleid')
qs1 = Sample.objects.all().filter(sampleid__in=selected)
if request.is_ajax():
allJson = request.GET["all"]
all = simplejson.loads(allJson)
button = int(all["button"])
sig_only = int(all["sig_only"])
norm = int(all["normalize"])
selectAll = int(all["selectAll"])
metaString = all["meta"]
### function to merge values on common keys
metaDict = simplejson.JSONDecoder(object_pairs_hook=multidict).decode(metaString)
### function to create a meta variable DataFrame
metaDF = catAlphaMetaDF(qs1, metaDict)
myList = metaDF['sampleid'].tolist()
mySet = list(set(myList))
### function to create a taxa DataFrame
taxaDF = taxaProfileDF(mySet)
### function to merge values on common keys
taxaString = all["taxa"]
### this taxaDict is from the dynatree (ajax call)
taxaDict = simplejson.JSONDecoder(object_pairs_hook=multidict).decode(taxaString)
# change dict if selectAll levels is on (avoids loading entire tree first)
if selectAll == 1:
taxaDict = {}
qs1 = Profile.objects.all().filter(sampleid__in=mySet).values_list('kingdomid', flat='True').distinct()
taxaDict['Kingdom'] = qs1
elif selectAll == 2:
taxaDict = {}
qs1 = Profile.objects.all().filter(sampleid__in=mySet).values_list('phylaid', flat='True').distinct()
taxaDict['Phyla'] = qs1
elif selectAll == 3:
taxaDict = {}
qs1 = Profile.objects.all().filter(sampleid__in=mySet).values_list('classid', flat='True').distinct()
taxaDict['Class'] = qs1
elif selectAll == 4:
taxaDict = {}
qs1 = Profile.objects.all().filter(sampleid__in=mySet).values_list('orderid', flat='True').distinct()
taxaDict['Order'] = qs1
elif selectAll == 5:
taxaDict = {}
qs1 = Profile.objects.all().filter(sampleid__in=mySet).values_list('familyid', flat='True').distinct()
taxaDict['Family'] = qs1
elif selectAll == 6:
taxaDict = {}
qs1 = Profile.objects.all().filter(sampleid__in=mySet).values_list('genusid', flat='True').distinct()
taxaDict['Genus'] = qs1
elif selectAll == 7:
taxaDict = {}
qs1 = Profile.objects.all().filter(sampleid__in=mySet).values_list('speciesid', flat='True').distinct()
taxaDict['Species'] = qs1
factor = 'none'
if norm == 1:
factor = 'none'
elif norm == 2:
factor = 'min'
elif norm == 3:
factor = '10th percentile'
elif norm == 4:
factor = '25th percentile'
elif norm == 5:
factor = 'median'
### function to normalize the number of sequence reads per sample
normDF = normalizeAlpha(taxaDF, taxaDict, mySet, factor)
finalDF = metaDF.merge(normDF, on='sampleid', how='outer')
finalDF[['count', 'rel_abund', 'rich', 'diversity']] = finalDF[['count', 'rel_abund', 'rich', 'diversity']].astype(float)
pd.set_option('display.max_rows', finalDF.shape[0], 'display.max_columns', finalDF.shape[1], 'display.width', 1000)
final_fieldList = []
for key in metaDict:
final_fieldList.append(key)
finalDict = {}
result = ""
seriesList = []
xAxisDict = {}
yAxisDict = {}
### group DataFrame by each taxa level selected
grouped1 = finalDF.groupby(['rank', 'taxa_name', 'taxa_id'])
equal_error = 'no'
### group DataFrame by each meta variable selected
for name1, group1 in grouped1:
trtList = []
valList = []
grouped2 = pd.DataFrame()
if button == 1:
grouped2 = group1.groupby(final_fieldList)['count']
elif button == 2:
grouped2 = group1.groupby(final_fieldList)['rel_abund']
elif button == 3:
grouped2 = group1.groupby(final_fieldList)['rich']
### for taxa with only 1 species all values will be '1' and cause an anova error
if group1['rich'].sum() == group1['rich'].count():
equal_error = 'yes'
elif button == 4:
grouped2 = group1.groupby(final_fieldList)['diversity']
for name2, group2 in grouped2:
if isinstance(name2, unicode):
trt = name2
else:
trt = ' & '.join(list(name2))
trtList.append(trt)
valList.append(list(group2.T))
### One-way ANOVA with some error checking
D = Anova1way()
if equal_error == 'no':
try:
D.run(valList, conditions_list=trtList)
anova_error = 'no'
except:
D['p'] = 1
anova_error = 'yes'
else:
D['p'] = 1
anova_error = 'yes'
### select only significant ANOVAs for output (graph & text area)
if sig_only == 1:
if D['p'] <= 0.05:
result = result + '===============================================\n'
result = result + 'Taxa level: ' + str(name1[0]) + '\n'
result = result + 'Taxa name: ' + str(name1[1]) + '\n'
if button == 1:
result = result + 'Dependent Variable: Sequence Reads' + '\n'
elif button == 2:
result = result + 'Dependent Variable: Relative Abundance' + '\n'
elif button == 3:
result = result + 'Dependent Variable: Species Richness' + '\n'
elif button == 4:
result = result + 'Dependent Variable: Shannon Diversity' + '\n'
indVar = ' x '.join(final_fieldList)
result = result + 'Independent Variable: ' + str(indVar) + '\n'
if equal_error == 'yes' or anova_error == 'yes':
result = result + 'Analysis cannot be performed...' + '\n'
else:
result = result + str(D) + '\n'
result = result + '===============================================\n'
result = result + '\n\n\n\n'
dataList = []
grouped2 = group1.groupby(final_fieldList).mean()
if button == 1:
dataList.extend(list(grouped2['count'].T))
elif button == 2:
dataList.extend(list(grouped2['rel_abund'].T))
elif button == 3:
dataList.extend(list(grouped2['rich'].T))
elif button == 4:
dataList.extend(list(grouped2['diversity'].T))
seriesDict = {}
seriesDict['name'] = name1
seriesDict['data'] = dataList
seriesList.append(seriesDict)
xTitle = {}
xTitle['text'] = indVar
xAxisDict['title'] = xTitle
xAxisDict['categories'] = trtList
yTitle = {}
if button == 1:
yTitle['text'] = 'Sequence Reads'
elif button == 2:
yTitle['text'] = 'Relative Abundance'
elif button == 3:
yTitle['text'] = 'Species Richness'
elif button == 4:
yTitle['text'] = 'Shannon Diversity'
yAxisDict['title'] = yTitle
### select all ANOVAs for output (graph & text area)
if sig_only == 0:
result = result + '===============================================\n'
result = result + 'Taxa level: ' + str(name1[0]) + '\n'
result = result + 'Taxa name: ' + str(name1[1]) + '\n'
if button == 1:
result = result + 'Dependent Variable: Sequence Reads' + '\n'
elif button == 2:
result = result + 'Dependent Variable: Relative Abundance' + '\n'
elif button == 3:
result = result + 'Dependent Variable: Species Richness' + '\n'
elif button == 4:
result = result + 'Dependent Variable: Shannon Diversity' + '\n'
indVar = ' x '.join(final_fieldList)
result = result + 'Independent Variable: ' + str(indVar) + '\n'
if equal_error == 'yes' or anova_error == 'yes':
result = result + 'Analysis cannot be performed...' + '\n'
else:
result = result + str(D) + '\n'
result = result + '===============================================\n'
result = result + '\n\n\n\n'
dataList = []
grouped2 = group1.groupby(final_fieldList).mean()
if button == 1:
dataList.extend(list(grouped2['count'].T))
elif button == 2:
dataList.extend(list(grouped2['rel_abund'].T))
elif button == 3:
dataList.extend(list(grouped2['rich'].T))
elif button == 4:
dataList.extend(list(grouped2['diversity'].T))
seriesDict = {}
seriesDict['name'] = name1
seriesDict['data'] = dataList
seriesList.append(seriesDict)
xTitle = {}
xTitle['text'] = indVar
xAxisDict['title'] = xTitle
xAxisDict['categories'] = trtList
yTitle = {}
if button == 1:
yTitle['text'] = 'Sequence Reads'
elif button == 2:
yTitle['text'] = 'Relative Abundance'
elif button == 3:
yTitle['text'] = 'Species Richness'
elif button == 4:
yTitle['text'] = 'Shannon Diversity'
yAxisDict['title'] = yTitle
finalDict['series'] = seriesList
finalDict['xAxis'] = xAxisDict
finalDict['yAxis'] = yAxisDict
finalDict['text'] = result
if not seriesList:
finalDict['empty'] = 0
else:
finalDict['empty'] = 1
finalDF.reset_index(drop=True, inplace=True)
res_table = finalDF.to_html(classes="table display")
res_table = res_table.replace('border="1"', 'border="0"')
finalDict['res_table'] = str(res_table)
res = simplejson.dumps(finalDict)
return HttpResponse(res, content_type='application/json')
def getCatBetaData(request):
samples = Sample.objects.all()
samples.query = pickle.loads(request.session['selected_samples'])
selected = samples.values_list('sampleid')
qs1 = Sample.objects.all().filter(sampleid__in=selected)
if request.is_ajax():
allJson = request.GET["all"]
all = simplejson.loads(allJson)
button = int(all["button"])
taxaLevel = int(all["taxa"])
distance = int(all["distance"])
norm = int(all["normalize"])
PC1 = all["PC1"]
PC2 = all["PC2"]
metaString = all["meta"]
metaDict = simplejson.JSONDecoder(object_pairs_hook=multidict).decode(metaString)
metaDF = catBetaMetaDF(qs1, metaDict)
myList = metaDF['sampleid'].tolist()
mySet = list(set(myList))
taxaDF = taxaProfileDF(mySet)
factor = 'none'
if norm == 1:
factor = 'none'
elif norm == 2:
factor = 'min'
elif norm == 3:
factor = '10th percentile'
elif norm == 4:
factor = '25th percentile'
elif norm == 5:
factor = 'median'
normDF = normalizeBeta(taxaDF, taxaLevel, mySet, factor)
finalDF = metaDF.merge(normDF, on='sampleid', how='outer')
pd.set_option('display.max_rows', finalDF.shape[0], 'display.max_columns', finalDF.shape[1], 'display.width', 1000)
fieldList = []
for key in metaDict:
fieldList.append(key)
matrixDF = pd.DataFrame()
if button == 1:
matrixDF = finalDF.pivot(index='taxaid', columns='sampleid', values='count')
elif button == 2:
matrixDF = finalDF.pivot(index='taxaid', columns='sampleid', values='rel_abund')
elif button == 3:
matrixDF = finalDF.pivot(index='taxaid', columns='sampleid', values='rich')
elif button == 4:
matrixDF = finalDF.pivot(index='taxaid', columns='sampleid', values='diversity')
datamtx = asarray(matrixDF[mySet].T)
numrows, numcols = shape(datamtx)
dists = np.zeros((numrows, numrows))
if distance == 1:
dist = pdist(datamtx, 'braycurtis')
dists = squareform(dist)
elif distance == 2:
dist = pdist(datamtx, 'canberra')
dists = squareform(dist)
elif distance == 3:
dist = pdist(datamtx, 'dice')
dists = squareform(dist)
elif distance == 4:
dist = pdist(datamtx, 'euclidean')
dists = squareform(dist)
elif distance == 5:
dist = pdist(datamtx, 'jaccard')
dists = squareform(dist)
eigvals, coordinates, proportion_explained = PCoA(dists)
numaxes = len(eigvals)
axesList = []
for i in range(numaxes):
j = i + 1
axesList.append('PC' + str(j))
valsDF = pd.DataFrame(eigvals, columns=['EigenVals'], index=axesList)
propDF = pd.DataFrame(proportion_explained, columns=['Variance Explained (R2)'], index=axesList)
eigenDF = valsDF.join(propDF)
metaDF.set_index('sampleid', drop=True, inplace=True)
pcoaDF = pd.DataFrame(coordinates, columns=axesList, index=mySet)
resultDF = metaDF.join(pcoaDF)
pd.set_option('display.max_rows', resultDF.shape[0], 'display.max_columns', resultDF.shape[1], 'display.width', 1000)
### create trtList that merges all categorical values
groupList = metaDF[fieldList].values.tolist()
trtList = []
for i in groupList:
trtList.append(':'.join(i))
### check to see if all samples are the same size
sizeList = []
grouped = resultDF.groupby(fieldList)
for name, group in grouped:
(row, col) = shape(group)
sizeList.append(row)
setSize = len(set(sizeList))
if setSize == 1:
try:
bigf, p = permanova_oneway(dists, trtList, 200)
except:
bigf = float('nan')
p = float('nan')
else:
bigf = float('nan')
p = float('nan')
finalDict = {}
seriesList = []
xAxisDict = {}
yAxisDict = {}
grouped = resultDF.groupby(fieldList)
for name, group in grouped:
dataList = group[[PC1, PC2]].values.tolist()
if isinstance(name, unicode):
trt = name
else:
trt = ' & '.join(list(name))
seriesDict = {}
seriesDict['name'] = trt
seriesDict['data'] = dataList
seriesList.append(seriesDict)
xTitle = {}
xTitle['text'] = PC1
xAxisDict['title'] = xTitle
yTitle = {}
yTitle['text'] = PC2
yAxisDict['title'] = yTitle
finalDict['series'] = seriesList
finalDict['xAxis'] = xAxisDict
finalDict['yAxis'] = yAxisDict
result = ""
result = result + '===============================================\n'
if taxaLevel == 1:
result = result + 'Taxa level: Kingdom' + '\n'
elif taxaLevel == 2:
result = result + 'Taxa level: Phyla' + '\n'
elif taxaLevel == 3:
result = result + 'Taxa level: Class' + '\n'
elif taxaLevel == 4:
result = result + 'Taxa level: Order' + '\n'
elif taxaLevel == 5:
result = result + 'Taxa level: Family' + '\n'
elif taxaLevel == 6:
result = result + 'Taxa level: Genus' + '\n'
elif taxaLevel == 7:
result = result + 'Taxa level: Species' + '\n'
if button == 1:
result = result + 'Dependent Variable: Sequence Reads' + '\n'
elif button == 2:
result = result + 'Dependent Variable: Relative Abundance' + '\n'
elif button == 3:
result = result + 'Dependent Variable: Species Richness' + '\n'
elif button == 4:
result = result + 'Dependent Variable: Shannon Diversity' + '\n'
indVar = ' x '.join(fieldList)
result = result + 'Independent Variable: ' + str(indVar) + '\n'
if distance == 1:
result = result + 'Distance score: Bray-Curtis' + '\n'
elif distance == 2:
result = result + 'Distance score: Canberra' + '\n'
elif distance == 3:
result = result + 'Distance score: Dice' + '\n'
elif distance == 4:
result = result + 'Distance score: Euclidean' + '\n'
elif distance == 5:
result = result + 'Distance score: Jaccard' + '\n'
if math.isnan(bigf):
result = result + '===============================================\n'
result = result + 'perMANOVA cannot be performed...' + '\n'
result = result + 'The current version requires all treatments to be of equal sample size.' + '\n'
else:
result = result + '===============================================\n'
result = result + 'perMANOVA results' + '\n'
result = result + 'f-value: ' + str(bigf) + '\n'
result = result + 'p-value: ' + str(p) + '\n'
result = result + '===============================================\n'
result = result + str(eigenDF) + '\n'
result = result + '===============================================\n'
result = result + '\n\n\n\n'
finalDict['text'] = result
resultDF.reset_index(drop=True, inplace=True)
res_table = resultDF.to_html(classes="table display")
res_table = res_table.replace('border="1"', 'border="0"')
finalDict['res_table'] = str(res_table)
nameList = list(metaDF['sample_name'])
distsDF = pd.DataFrame(dists, columns=nameList, index=nameList)
dist_table = distsDF.to_html(classes="table display")
dist_table = dist_table.replace('border="1"', 'border="0"')
finalDict['dist_table'] = str(dist_table)
res = simplejson.dumps(finalDict)
return HttpResponse(res, content_type='application/json')
def getQuantAlphaData(request):
samples = Sample.objects.all()
samples.query = pickle.loads(request.session['selected_samples'])
selected = samples.values_list('sampleid')
qs1 = Sample.objects.all().filter(sampleid__in=selected)
if request.is_ajax():
allJson = request.GET["all"]
all = simplejson.loads(allJson)
button = int(all["button"])
sig_only = int(all["sig_only"])
norm = int(all["normalize"])
taxaString = all["taxa"]
taxaDict = simplejson.JSONDecoder(
object_pairs_hook=multidict).decode(taxaString)
metaString = all["meta"]
metaDict = simplejson.JSONDecoder(
object_pairs_hook=multidict).decode(metaString)
metaDF = quantAlphaMetaDF(qs1, metaDict)
myList = metaDF['sampleid'].tolist()
mySet = list(set(myList))
taxaDF = taxaProfileDF(mySet)
factor = 'none'
if norm == 1:
factor = 'none'
elif norm == 2:
factor = 'min'
elif norm == 3:
factor = '10th percentile'
elif norm == 4:
factor = '25th percentile'
elif norm == 5:
factor = 'median'
final_fieldList = []
for key in metaDict:
final_fieldList.append(metaDict[key])
normDF = normalizeAlpha(taxaDF, taxaDict, mySet, factor)
finalDF = metaDF.merge(normDF, on='sampleid', how='outer')
finalDF[[
final_fieldList[0], 'count', 'rel_abund', 'rich', 'diversity'
]] = finalDF[[
final_fieldList[0], 'count', 'rel_abund', 'rich', 'diversity'
]].astype(float)
pd.set_option('display.max_rows', finalDF.shape[0],
'display.max_columns', finalDF.shape[1], 'display.width',
1000)
finalDict = {}
seriesList = []
xAxisDict = {}
yAxisDict = {}
grouped1 = finalDF.groupby(['rank', 'taxa_name', 'taxa_id'])
for name1, group1 in grouped1:
dataList = []
x = []
y = []
if button == 1:
dataList = group1[[final_fieldList[0],
'count']].values.tolist()
x = group1[final_fieldList[0]].values.tolist()
y = group1['count'].values.tolist()
elif button == 2:
dataList = group1[[final_fieldList[0],
'rel_abund']].values.tolist()
x = group1[final_fieldList[0]].values.tolist()
y = group1['rel_abund'].values.tolist()
elif button == 3:
dataList = group1[[final_fieldList[0], 'rich']].values.tolist()
x = group1[final_fieldList[0]].values.tolist()
y = group1['rich'].values.tolist()
elif button == 4:
dataList = group1[[final_fieldList[0],
'diversity']].values.tolist()
x = group1[final_fieldList[0]].values.tolist()
y = group1['diversity'].values.tolist()
if max(x) == min(x):
stop = 0
else:
stop = 1
slope, intercept, r_value, p_value, std_err = stats.linregress(
x, y)
p_value = "%0.3f" % p_value
r_square = r_value * r_value
r_square = "%0.4f" % r_square
min_y = slope * min(x) + intercept
max_y = slope * max(x) + intercept
slope = "%.3E" % slope
intercept = "%.3E" % intercept
regrList = []
regrList.append([min(x), min_y])
regrList.append([max(x), max_y])
if sig_only == 0:
seriesDict = {}
seriesDict['type'] = 'scatter'
seriesDict['name'] = name1
seriesDict['data'] = dataList
seriesList.append(seriesDict)
if stop == 0:
regDict = {}
elif stop == 1:
regrDict = {}
regrDict['type'] = 'line'
name2 = list(name1)
temp = 'R2: ' + str(r_square) + '; p-value: ' + str(
p_value) + '<br>' + '(y = ' + str(
slope) + 'x' + ' + ' + str(intercept)
print temp
name2.append(temp)
print name2
regrDict['name'] = name2
regrDict['data'] = regrList
seriesList.append(regrDict)
if sig_only == 1:
if p_value <= 0.05:
seriesDict = {}
seriesDict['type'] = 'scatter'
name2 = list(name1)
temp = 'R2: ' + str(r_square) + '; p-value: ' + str(
p_value) + '<br>' + '(y = ' + str(
slope) + 'x' + ' + ' + str(intercept)
name2.append(temp)
seriesDict['name'] = name2
seriesDict['data'] = dataList
seriesList.append(seriesDict)
regrDict = {}
regrDict['type'] = 'line'
regrDict['name'] = name1
regrDict['data'] = regrList
seriesList.append(regrDict)
xTitle = {}
xTitle['text'] = final_fieldList[0]
xAxisDict['title'] = xTitle
yTitle = {}
if button == 1:
yTitle['text'] = 'Sequence Reads'
elif button == 2:
yTitle['text'] = 'Relative Abundance'
elif button == 3:
yTitle['text'] = 'Species Richness'
elif button == 4:
yTitle['text'] = 'Shannon Diversity'
yAxisDict['title'] = yTitle
finalDict['series'] = seriesList
finalDict['xAxis'] = xAxisDict
finalDict['yAxis'] = yAxisDict
if not seriesList:
finalDict['empty'] = 0
else:
finalDict['empty'] = 1
finalDF.reset_index(drop=True, inplace=True)
res_table = finalDF.to_html(classes="table display")
res_table = res_table.replace('border="1"', 'border="0"')
finalDict['res_table'] = str(res_table)
res = simplejson.dumps(finalDict)
return HttpResponse(res, content_type='application/json')
def getCatAlphaData(request):
### get sample list from cookie
samples = Sample.objects.all()
samples.query = pickle.loads(request.session['selected_samples'])
selected = samples.values_list('sampleid')
qs1 = Sample.objects.all().filter(sampleid__in=selected)
if request.is_ajax():
allJson = request.GET["all"]
all = simplejson.loads(allJson)
button = int(all["button"])
sig_only = int(all["sig_only"])
norm = int(all["normalize"])
selectAll = int(all["selectAll"])
metaString = all["meta"]
### function to merge values on common keys
metaDict = simplejson.JSONDecoder(
object_pairs_hook=multidict).decode(metaString)
### function to create a meta variable DataFrame
metaDF = catAlphaMetaDF(qs1, metaDict)
myList = metaDF['sampleid'].tolist()
mySet = list(set(myList))
### function to create a taxa DataFrame
taxaDF = taxaProfileDF(mySet)
### function to merge values on common keys
taxaString = all["taxa"]
### this taxaDict is from the dynatree (ajax call)
taxaDict = simplejson.JSONDecoder(
object_pairs_hook=multidict).decode(taxaString)
# change dict if selectAll levels is on (avoids loading entire tree first)
if selectAll == 1:
taxaDict = {}
qs1 = Profile.objects.all().filter(sampleid__in=mySet).values_list(
'kingdomid', flat='True').distinct()
taxaDict['Kingdom'] = qs1
elif selectAll == 2:
taxaDict = {}
qs1 = Profile.objects.all().filter(sampleid__in=mySet).values_list(
'phylaid', flat='True').distinct()
taxaDict['Phyla'] = qs1
elif selectAll == 3:
taxaDict = {}
qs1 = Profile.objects.all().filter(sampleid__in=mySet).values_list(
'classid', flat='True').distinct()
taxaDict['Class'] = qs1
elif selectAll == 4:
taxaDict = {}
qs1 = Profile.objects.all().filter(sampleid__in=mySet).values_list(
'orderid', flat='True').distinct()
taxaDict['Order'] = qs1
elif selectAll == 5:
taxaDict = {}
qs1 = Profile.objects.all().filter(sampleid__in=mySet).values_list(
'familyid', flat='True').distinct()
taxaDict['Family'] = qs1
elif selectAll == 6:
taxaDict = {}
qs1 = Profile.objects.all().filter(sampleid__in=mySet).values_list(
'genusid', flat='True').distinct()
taxaDict['Genus'] = qs1
elif selectAll == 7:
taxaDict = {}
qs1 = Profile.objects.all().filter(sampleid__in=mySet).values_list(
'speciesid', flat='True').distinct()
taxaDict['Species'] = qs1
factor = 'none'
if norm == 1:
factor = 'none'
elif norm == 2:
factor = 'min'
elif norm == 3:
factor = '10th percentile'
elif norm == 4:
factor = '25th percentile'
elif norm == 5:
factor = 'median'
### function to normalize the number of sequence reads per sample
normDF = normalizeAlpha(taxaDF, taxaDict, mySet, factor)
finalDF = metaDF.merge(normDF, on='sampleid', how='outer')
finalDF[['count', 'rel_abund', 'rich', 'diversity'
]] = finalDF[['count', 'rel_abund', 'rich',
'diversity']].astype(float)
pd.set_option('display.max_rows', finalDF.shape[0],
'display.max_columns', finalDF.shape[1], 'display.width',
1000)
final_fieldList = []
for key in metaDict:
final_fieldList.append(key)
finalDict = {}
result = ""
seriesList = []
xAxisDict = {}
yAxisDict = {}
### group DataFrame by each taxa level selected
grouped1 = finalDF.groupby(['rank', 'taxa_name', 'taxa_id'])
equal_error = 'no'
### group DataFrame by each meta variable selected
for name1, group1 in grouped1:
trtList = []
valList = []
grouped2 = pd.DataFrame()
if button == 1:
grouped2 = group1.groupby(final_fieldList)['count']
elif button == 2:
grouped2 = group1.groupby(final_fieldList)['rel_abund']
elif button == 3:
grouped2 = group1.groupby(final_fieldList)['rich']
### for taxa with only 1 species all values will be '1' and cause an anova error
if group1['rich'].sum() == group1['rich'].count():
equal_error = 'yes'
elif button == 4:
grouped2 = group1.groupby(final_fieldList)['diversity']
for name2, group2 in grouped2:
if isinstance(name2, unicode):
trt = name2
else:
trt = ' & '.join(list(name2))
trtList.append(trt)
valList.append(list(group2.T))
### One-way ANOVA with some error checking
D = Anova1way()
if equal_error == 'no':
try:
D.run(valList, conditions_list=trtList)
anova_error = 'no'
except:
D['p'] = 1
anova_error = 'yes'
else:
D['p'] = 1
anova_error = 'yes'
### select only significant ANOVAs for output (graph & text area)
if sig_only == 1:
if D['p'] <= 0.05:
result = result + '===============================================\n'
result = result + 'Taxa level: ' + str(name1[0]) + '\n'
result = result + 'Taxa name: ' + str(name1[1]) + '\n'
if button == 1:
result = result + 'Dependent Variable: Sequence Reads' + '\n'
elif button == 2:
result = result + 'Dependent Variable: Relative Abundance' + '\n'
elif button == 3:
result = result + 'Dependent Variable: Species Richness' + '\n'
elif button == 4:
result = result + 'Dependent Variable: Shannon Diversity' + '\n'
indVar = ' x '.join(final_fieldList)
result = result + 'Independent Variable: ' + str(
indVar) + '\n'
if equal_error == 'yes' or anova_error == 'yes':
result = result + 'Analysis cannot be performed...' + '\n'
else:
result = result + str(D) + '\n'
result = result + '===============================================\n'
result = result + '\n\n\n\n'
dataList = []
grouped2 = group1.groupby(final_fieldList).mean()
if button == 1:
dataList.extend(list(grouped2['count'].T))
elif button == 2:
dataList.extend(list(grouped2['rel_abund'].T))
elif button == 3:
dataList.extend(list(grouped2['rich'].T))
elif button == 4:
dataList.extend(list(grouped2['diversity'].T))
seriesDict = {}
seriesDict['name'] = name1
seriesDict['data'] = dataList
seriesList.append(seriesDict)
xTitle = {}
xTitle['text'] = indVar
xAxisDict['title'] = xTitle
xAxisDict['categories'] = trtList
yTitle = {}
if button == 1:
yTitle['text'] = 'Sequence Reads'
elif button == 2:
yTitle['text'] = 'Relative Abundance'
elif button == 3:
yTitle['text'] = 'Species Richness'
elif button == 4:
yTitle['text'] = 'Shannon Diversity'
yAxisDict['title'] = yTitle
### select all ANOVAs for output (graph & text area)
if sig_only == 0:
result = result + '===============================================\n'
result = result + 'Taxa level: ' + str(name1[0]) + '\n'
result = result + 'Taxa name: ' + str(name1[1]) + '\n'
if button == 1:
result = result + 'Dependent Variable: Sequence Reads' + '\n'
elif button == 2:
result = result + 'Dependent Variable: Relative Abundance' + '\n'
elif button == 3:
result = result + 'Dependent Variable: Species Richness' + '\n'
elif button == 4:
result = result + 'Dependent Variable: Shannon Diversity' + '\n'
indVar = ' x '.join(final_fieldList)
result = result + 'Independent Variable: ' + str(indVar) + '\n'
if equal_error == 'yes' or anova_error == 'yes':
result = result + 'Analysis cannot be performed...' + '\n'
else:
result = result + str(D) + '\n'
result = result + '===============================================\n'
result = result + '\n\n\n\n'
dataList = []
grouped2 = group1.groupby(final_fieldList).mean()
if button == 1:
dataList.extend(list(grouped2['count'].T))
elif button == 2:
dataList.extend(list(grouped2['rel_abund'].T))
elif button == 3:
dataList.extend(list(grouped2['rich'].T))
elif button == 4:
dataList.extend(list(grouped2['diversity'].T))
seriesDict = {}
seriesDict['name'] = name1
seriesDict['data'] = dataList
seriesList.append(seriesDict)
xTitle = {}
xTitle['text'] = indVar
xAxisDict['title'] = xTitle
xAxisDict['categories'] = trtList
yTitle = {}
if button == 1:
yTitle['text'] = 'Sequence Reads'
elif button == 2:
yTitle['text'] = 'Relative Abundance'
elif button == 3:
yTitle['text'] = 'Species Richness'
elif button == 4:
yTitle['text'] = 'Shannon Diversity'
yAxisDict['title'] = yTitle
finalDict['series'] = seriesList
finalDict['xAxis'] = xAxisDict
finalDict['yAxis'] = yAxisDict
finalDict['text'] = result
if not seriesList:
finalDict['empty'] = 0
else:
finalDict['empty'] = 1
finalDF.reset_index(drop=True, inplace=True)
res_table = finalDF.to_html(classes="table display")
res_table = res_table.replace('border="1"', 'border="0"')
finalDict['res_table'] = str(res_table)
res = simplejson.dumps(finalDict)
return HttpResponse(res, content_type='application/json')
def getQuantBetaData(request):
samples = Sample.objects.all()
samples.query = pickle.loads(request.session['selected_samples'])
selected = samples.values_list('sampleid')
qs1 = Sample.objects.all().filter(sampleid__in=selected)
if request.is_ajax():
allJson = request.GET["all"]
all = simplejson.loads(allJson)
button = int(all["button"])
taxaLevel = int(all["taxa"])
distance = int(all["distance"])
norm = int(all["normalize"])
PC1 = all["PC1"]
metaString = all["meta"]
metaDict = simplejson.JSONDecoder(
object_pairs_hook=multidict).decode(metaString)
metaDF = quantBetaMetaDF(qs1, metaDict)
myList = metaDF['sampleid'].tolist()
mySet = list(set(myList))
taxaDF = taxaProfileDF(mySet)
factor = 'none'
if norm == 1:
factor = 'none'
elif norm == 2:
factor = 'min'
elif norm == 3:
factor = '10th percentile'
elif norm == 4:
factor = '25th percentile'
elif norm == 5:
factor = 'median'
normDF = normalizeBeta(taxaDF, taxaLevel, mySet, factor)
finalDF = metaDF.merge(normDF, on='sampleid', how='outer')
pd.set_option('display.max_rows', finalDF.shape[0],
'display.max_columns', finalDF.shape[1], 'display.width',
1000)
fieldList = []
for key in metaDict:
fieldList.append(metaDict[key])
matrixDF = pd.DataFrame()
if button == 1:
matrixDF = finalDF.pivot(index='taxaid',
columns='sampleid',
values='count')
elif button == 2:
matrixDF = finalDF.pivot(index='taxaid',
columns='sampleid',
values='rel_abund')
elif button == 3:
matrixDF = finalDF.pivot(index='taxaid',
columns='sampleid',
values='rich')
elif button == 4:
matrixDF = finalDF.pivot(index='taxaid',
columns='sampleid',
values='diversity')
datamtx = asarray(matrixDF[mySet].T)
numrows, numcols = shape(datamtx)
dists = zeros((numrows, numrows))
if distance == 1:
dist = pdist(datamtx, 'braycurtis')
dists = squareform(dist)
elif distance == 2:
dist = pdist(datamtx, 'canberra')
dists = squareform(dist)
elif distance == 3:
dist = pdist(datamtx, 'dice')
dists = squareform(dist)
elif distance == 4:
dist = pdist(datamtx, 'euclidean')
dists = squareform(dist)
elif distance == 5:
dist = pdist(datamtx, 'jaccard')
dists = squareform(dist)
eigvals, coordinates, proportion_explained = PCoA(dists)
numaxes = len(eigvals)
axesList = []
for i in range(numaxes):
j = i + 1
axesList.append('PC' + str(j))
valsDF = pd.DataFrame(eigvals, columns=['EigenVals'], index=axesList)
propDF = pd.DataFrame(proportion_explained,
columns=['Variance Explained (R2)'],
index=axesList)
eigenDF = valsDF.join(propDF)
metaDF.set_index('sampleid', drop=True, inplace=True)
pcoaDF = pd.DataFrame(coordinates, columns=axesList, index=mySet)
resultDF = metaDF.join(pcoaDF)
pd.set_option('display.max_rows', resultDF.shape[0],
'display.max_columns', resultDF.shape[1],
'display.width', 1000)
finalDict = {}
seriesList = []
xAxisDict = {}
yAxisDict = {}
dataList = resultDF[[PC1, fieldList[0]]].values.tolist()
seriesDict = {}
seriesDict['type'] = 'scatter'
seriesDict['name'] = fieldList
seriesDict['data'] = dataList
seriesList.append(seriesDict)
x = resultDF[PC1].values.tolist()
y = resultDF[fieldList[0]].values.tolist()
if max(x) == min(x):
stop = 0
else:
stop = 1
slope, intercept, r_value, p_value, std_err = stats.linregress(
x, y)
p_value = "%0.3f" % p_value
r_square = r_value * r_value
r_square = "%0.4f" % r_square
min_y = slope * min(x) + intercept
max_y = slope * max(x) + intercept
slope = "%.3E" % slope
intercept = "%.3E" % intercept
regrList = []
regrList.append([min(x), min_y])
regrList.append([max(x), max_y])
if stop == 0:
regDict = {}
elif stop == 1:
regrDict = {}
regrDict['type'] = 'line'
regrDict['name'] = 'R2: ' + str(
r_square) + '; p-value: ' + str(
p_value) + '<br>' + '(y = ' + str(
slope) + 'x' + ' + ' + str(intercept) + ')'
regrDict['data'] = regrList
seriesList.append(regrDict)
xTitle = {}
xTitle['text'] = PC1
xAxisDict['title'] = xTitle
yTitle = {}
yTitle['text'] = fieldList[0]
yAxisDict['title'] = yTitle
finalDict['series'] = seriesList
finalDict['xAxis'] = xAxisDict
finalDict['yAxis'] = yAxisDict
result = ""
result = result + '===============================================\n'
if taxaLevel == 1:
result = result + 'Taxa level: Kingdom' + '\n'
elif taxaLevel == 2:
result = result + 'Taxa level: Phyla' + '\n'
elif taxaLevel == 3:
result = result + 'Taxa level: Class' + '\n'
elif taxaLevel == 4:
result = result + 'Taxa level: Order' + '\n'
elif taxaLevel == 5:
result = result + 'Taxa level: Family' + '\n'
elif taxaLevel == 6:
result = result + 'Taxa level: Genus' + '\n'
elif taxaLevel == 7:
result = result + 'Taxa level: Species' + '\n'
if button == 1:
result = result + 'Dependent Variable: Sequence Reads' + '\n'
elif button == 2:
result = result + 'Dependent Variable: Relative Abundance' + '\n'
elif button == 3:
result = result + 'Dependent Variable: Species Richness' + '\n'
elif button == 4:
result = result + 'Dependent Variable: Shannon Diversity' + '\n'
result = result + 'Independent Variable: ' + str(fieldList[0]) + '\n'
if distance == 1:
result = result + 'Distance score: Bray-Curtis' + '\n'
elif distance == 2:
result = result + 'Distance score: Canberra' + '\n'
elif distance == 3:
result = result + 'Distance score: Dice' + '\n'
elif distance == 4:
result = result + 'Distance score: Euclidean' + '\n'
elif distance == 5:
result = result + 'Distance score: Jaccard' + '\n'
result = result + '===============================================\n'
result = result + str(eigenDF) + '\n'
result = result + '===============================================\n'
result = result + '\n\n\n\n'
finalDict['text'] = result
resultDF.reset_index(drop=True, inplace=True)
res_table = resultDF.to_html(classes="table display")
res_table = res_table.replace('border="1"', 'border="0"')
finalDict['res_table'] = str(res_table)
nameList = list(metaDF['sample_name'])
distsDF = pd.DataFrame(dists, columns=nameList, index=nameList)
dist_table = distsDF.to_html(classes="table display")
dist_table = dist_table.replace('border="1"', 'border="0"')
finalDict['dist_table'] = str(dist_table)
res = simplejson.dumps(finalDict)
return HttpResponse(res, content_type='application/json')
def getCatBetaData(request):
samples = Sample.objects.all()
samples.query = pickle.loads(request.session['selected_samples'])
selected = samples.values_list('sampleid')
qs1 = Sample.objects.all().filter(sampleid__in=selected)
if request.is_ajax():
allJson = request.GET["all"]
all = simplejson.loads(allJson)
button = int(all["button"])
taxaLevel = int(all["taxa"])
distance = int(all["distance"])
norm = int(all["normalize"])
PC1 = all["PC1"]
PC2 = all["PC2"]
metaString = all["meta"]
metaDict = simplejson.JSONDecoder(
object_pairs_hook=multidict).decode(metaString)
metaDF = catBetaMetaDF(qs1, metaDict)
myList = metaDF['sampleid'].tolist()
mySet = list(set(myList))
taxaDF = taxaProfileDF(mySet)
factor = 'none'
if norm == 1:
factor = 'none'
elif norm == 2:
factor = 'min'
elif norm == 3:
factor = '10th percentile'
elif norm == 4:
factor = '25th percentile'
elif norm == 5:
factor = 'median'
normDF = normalizeBeta(taxaDF, taxaLevel, mySet, factor)
finalDF = metaDF.merge(normDF, on='sampleid', how='outer')
pd.set_option('display.max_rows', finalDF.shape[0],
'display.max_columns', finalDF.shape[1], 'display.width',
1000)
fieldList = []
for key in metaDict:
fieldList.append(key)
matrixDF = pd.DataFrame()
if button == 1:
matrixDF = finalDF.pivot(index='taxaid',
columns='sampleid',
values='count')
elif button == 2:
matrixDF = finalDF.pivot(index='taxaid',
columns='sampleid',
values='rel_abund')
elif button == 3:
matrixDF = finalDF.pivot(index='taxaid',
columns='sampleid',
values='rich')
elif button == 4:
matrixDF = finalDF.pivot(index='taxaid',
columns='sampleid',
values='diversity')
datamtx = asarray(matrixDF[mySet].T)
numrows, numcols = shape(datamtx)
dists = np.zeros((numrows, numrows))
if distance == 1:
dist = pdist(datamtx, 'braycurtis')
dists = squareform(dist)
elif distance == 2:
dist = pdist(datamtx, 'canberra')
dists = squareform(dist)
elif distance == 3:
dist = pdist(datamtx, 'dice')
dists = squareform(dist)
elif distance == 4:
dist = pdist(datamtx, 'euclidean')
dists = squareform(dist)
elif distance == 5:
dist = pdist(datamtx, 'jaccard')
dists = squareform(dist)
eigvals, coordinates, proportion_explained = PCoA(dists)
numaxes = len(eigvals)
axesList = []
for i in range(numaxes):
j = i + 1
axesList.append('PC' + str(j))
valsDF = pd.DataFrame(eigvals, columns=['EigenVals'], index=axesList)
propDF = pd.DataFrame(proportion_explained,
columns=['Variance Explained (R2)'],
index=axesList)
eigenDF = valsDF.join(propDF)
metaDF.set_index('sampleid', drop=True, inplace=True)
pcoaDF = pd.DataFrame(coordinates, columns=axesList, index=mySet)
resultDF = metaDF.join(pcoaDF)
pd.set_option('display.max_rows', resultDF.shape[0],
'display.max_columns', resultDF.shape[1],
'display.width', 1000)
### create trtList that merges all categorical values
groupList = metaDF[fieldList].values.tolist()
trtList = []
for i in groupList:
trtList.append(':'.join(i))
### check to see if all samples are the same size
sizeList = []
grouped = resultDF.groupby(fieldList)
for name, group in grouped:
(row, col) = shape(group)
sizeList.append(row)
setSize = len(set(sizeList))
if setSize == 1:
try:
bigf, p = permanova_oneway(dists, trtList, 200)
except:
bigf = float('nan')
p = float('nan')
else:
bigf = float('nan')
p = float('nan')
finalDict = {}
seriesList = []
xAxisDict = {}
yAxisDict = {}
grouped = resultDF.groupby(fieldList)
for name, group in grouped:
dataList = group[[PC1, PC2]].values.tolist()
if isinstance(name, unicode):
trt = name
else:
trt = ' & '.join(list(name))
seriesDict = {}
seriesDict['name'] = trt
seriesDict['data'] = dataList
seriesList.append(seriesDict)
xTitle = {}
xTitle['text'] = PC1
xAxisDict['title'] = xTitle
yTitle = {}
yTitle['text'] = PC2
yAxisDict['title'] = yTitle
finalDict['series'] = seriesList
finalDict['xAxis'] = xAxisDict
finalDict['yAxis'] = yAxisDict
result = ""
result = result + '===============================================\n'
if taxaLevel == 1:
result = result + 'Taxa level: Kingdom' + '\n'
elif taxaLevel == 2:
result = result + 'Taxa level: Phyla' + '\n'
elif taxaLevel == 3:
result = result + 'Taxa level: Class' + '\n'
elif taxaLevel == 4:
result = result + 'Taxa level: Order' + '\n'
elif taxaLevel == 5:
result = result + 'Taxa level: Family' + '\n'
elif taxaLevel == 6:
result = result + 'Taxa level: Genus' + '\n'
elif taxaLevel == 7:
result = result + 'Taxa level: Species' + '\n'
if button == 1:
result = result + 'Dependent Variable: Sequence Reads' + '\n'
elif button == 2:
result = result + 'Dependent Variable: Relative Abundance' + '\n'
elif button == 3:
result = result + 'Dependent Variable: Species Richness' + '\n'
elif button == 4:
result = result + 'Dependent Variable: Shannon Diversity' + '\n'
indVar = ' x '.join(fieldList)
result = result + 'Independent Variable: ' + str(indVar) + '\n'
if distance == 1:
result = result + 'Distance score: Bray-Curtis' + '\n'
elif distance == 2:
result = result + 'Distance score: Canberra' + '\n'
elif distance == 3:
result = result + 'Distance score: Dice' + '\n'
elif distance == 4:
result = result + 'Distance score: Euclidean' + '\n'
elif distance == 5:
result = result + 'Distance score: Jaccard' + '\n'
if math.isnan(bigf):
result = result + '===============================================\n'
result = result + 'perMANOVA cannot be performed...' + '\n'
result = result + 'The current version requires all treatments to be of equal sample size.' + '\n'
else:
result = result + '===============================================\n'
result = result + 'perMANOVA results' + '\n'
result = result + 'f-value: ' + str(bigf) + '\n'
result = result + 'p-value: ' + str(p) + '\n'
result = result + '===============================================\n'
result = result + str(eigenDF) + '\n'
result = result + '===============================================\n'
result = result + '\n\n\n\n'
finalDict['text'] = result
resultDF.reset_index(drop=True, inplace=True)
res_table = resultDF.to_html(classes="table display")
res_table = res_table.replace('border="1"', 'border="0"')
finalDict['res_table'] = str(res_table)
nameList = list(metaDF['sample_name'])
distsDF = pd.DataFrame(dists, columns=nameList, index=nameList)
dist_table = distsDF.to_html(classes="table display")
dist_table = dist_table.replace('border="1"', 'border="0"')
finalDict['dist_table'] = str(dist_table)
res = simplejson.dumps(finalDict)
return HttpResponse(res, content_type='application/json')
