def load_genusData(gfilename):
#read csv file
csv_list, nrows, ncols = csvReader.csv_reader(gfilename)
x_data = csv_list[0][2:]
genus_dic = {}
table_sum = [0 for i in range(len(x_data))]
bottom_data = {}
bottom_data[0] = [0 for i in range(len(x_data))]
for row_num in range(1, nrows):
current_genus = csv_list[row_num][0]
current_type = csv_list[row_num][1]
genus_dic[row_num] = {}
genus_dic[row_num]['type'] = current_type
genus_dic[row_num]['rate'] = csv_list[row_num][2:]
genus_dic[row_num]['name'] = current_genus
table_sum = [i + j for i, j in zip(table_sum, csv_list[row_num][2:])]
bottom_data[row_num] = table_sum
#conver to percent
for row_num in range(1, nrows):
genus_dic[row_num]['rate'] = [
i / j * 100 for i, j in zip(genus_dic[row_num]['rate'], table_sum)
]
bottom_data[row_num] = [
i / j * 100 for i, j in zip(bottom_data[row_num], table_sum)
]
return nrows, x_data, genus_dic, bottom_data
def load_bacterium(genus_filename, group_filename):
#read csv file
csv_list, nrows, ncols = csvReader.csv_reader(genus_filename)
#read group file
group, group_names = load_groupData(group_filename)
bacterium = {}
sample_data = []
group_id = []
sample_names = []
for col_num in range(2, ncols):
sample_name = csv_list[0][col_num]
group_id.append(group[sample_name])
sample_names.append(sample_name)
data = []
for row_num in range(1, nrows):
data.append(csv_list[row_num][col_num])
sample_data.append(data)
bacterium['data'] = np.asarray(sample_data)
bacterium['sample'] = np.asarray(sample_names)
bacterium['group_id'] = np.asarray(group_id)
bacterium['group_names'] = np.asarray(group_names)
return bacterium
def get_genusSize(gfilename):
csv_list, nrows, ncols = csvReader.csv_reader(gfilename)
genus_size = {}
for row_num in range(1, nrows):
current_genus = csv_list[row_num][0]
genus_size[current_genus] = csv_list[row_num][2:]
return genus_size
def load_groupData(group_filename):
#read group file
csv_list, nrows, ncols = csvReader.csv_reader(group_filename)
group = {}
group_names = []
for n in range(nrows):
if not csv_list[n][1].lower() in group_names:
group_names.append(csv_list[n][1].lower())
group[csv_list[n][0]] = group_names.index(csv_list[n][1].lower())
return group, group_names
def kinship(dir):
kinList = [[], []]
for s in range(dir):
csvFile = getCSV(rootdir, s)
famList = csv_reader(csvFile)
#in famList, fathers are indexed in 0 and sons in 1
#print famList
for rows in famList:
imFather = folderFind(rootdir, s, rows[0])
imSon = folderFind(rootdir, s, rows[1])
if len(imFather) > 0 and len(imSon) > 0:
if len(imFather) > len(imSon):
for i in range(len(imFather) - len(imSon)):
imSon.append(imSon[0])
elif len(imSon) > len(imFather):
for i in range(len(imSon) - len(imFather)):
imFather.append(imFather[0])
kinList[0].append(imFather)
kinList[1].append(imSon)
#print len(kinList[1])
#print len(kinList[0])
dads = [val for sublist in kinList[0] for val in sublist] #Flatten lists
sons = [val for sublist in kinList[1] for val in sublist] #Flatten lists
print len(dads)
#print len(sons)
kinList = [[], []] #Clear list to copy unidimentional list
for n in range(len(dads)):
kinList[0].append(dads[n])
kinList[1].append(sons[n])
return kinList
def load_speciesData(sfilename, gfilename):
#read csv file
csv_list, nrows, ncols = csvReader.csv_reader(sfilename)
genus_size = get_genusSize(gfilename)
x_data = csv_list[0][2:]
species_dic = {}
bottom_data = {}
for row_num in range(1, nrows):
current_genus = csv_list[row_num][0]
current_species = csv_list[row_num][1]
if (species_dic.get(current_genus) == None):
n = 1
species_dic[current_genus] = {}
bottom_data[current_genus] = {}
bottom_data[current_genus][0] = [0 for i in range(len(x_data))]
n = n - 1
n = n + 1
species_dic[current_genus][n] = {}
if genus_size.get(current_genus) != None:
#size of species_bar is absolute value! not relative!
species_dic[current_genus][n]['rate'] = [
i * j for i, j in zip(csv_list[row_num][2:],
genus_size[current_genus])
]
species_dic[current_genus][n]['name'] = current_species
bottom_data[current_genus][n] = [
i + j for i, j in zip(bottom_data[current_genus][n - 1],
species_dic[current_genus][n]['rate'])
]
return nrows, x_data, species_dic, bottom_data
def run(filename, group_filename, checked) :
#read csv file
csv_list, nrows, ncols = csvReader.csv_reader(filename)
x_data = csv_list[0][2:]
#load group file
group, group_names = load_groupData(group_filename)
global group1
global group2
group1 = []
group2 = []
#index of group in x_data
group1_ind = []
group2_ind = []
#separate group1 and group2
for n in range(len(group)) :
group1_id = group_names.index(checked[0])
group2_id = group_names.index(checked[1])
if group.values()[n] == group1_id :
group1.append(group.keys()[n])
group1_ind.append(x_data.index(group.keys()[n]))
elif group.values()[n] == group2_id :
group2.append(group.keys()[n])
group2_ind.append(x_data.index(group.keys()[n]))
filtered = {}
filtered['data'] = []
filtered['genus'] = []
filtered['pvalue'] = []
#extract data which have low pvalue
for row_num in range(1,nrows):
current_genus = csv_list[row_num][0]
current_type = csv_list[row_num][1]
test_group1 = []
test_group2 = []
for i in group1_ind :
test_group1.append(csv_list[row_num][2+i])
for i in group2_ind :
test_group2.append(csv_list[row_num][2+i])
# calculate pvalue
pvalue = sci.ttest_ind(test_group1, test_group2, equal_var=True)[1]
if pvalue < 0.05 :
filtered['data'].append(test_group1+test_group2)
filtered['genus'].append(current_genus+"("+str(current_type)[0]+")")
filtered['pvalue'].append(round(pvalue,4))
d3 = bd.draw_graph(group1, group2, checked,
x_label = [i for i in range(len(group1+group2))],
y_label=[i for i in range(len(filtered['data']))],
pvalue_label = filtered['pvalue'],
fit_data = filtered['data'],
genus_data = filtered['genus'],
title = "Original dataset",
x = [i for i in range(len(group1+group2))])
d3.draw()
if len(filtered['data']) > 0 :
biclustering(filtered, checked)
else :
print "no data"
plt.show()
def run(genus_filename, group_filename):
#load genus file
csv_list, nrows, ncols = csvReader.csv_reader(genus_filename)
#load bacterium data
bacterium = load_bacterium(genus_filename, group_filename)
X = bacterium['data']
y = bacterium['group_id']
group_names = bacterium['group_names']
n_components = 2
pca = PCA(n_components=n_components)
X_pca = pca.fit_transform(X)
colors = ['navy', 'turquoise', 'darkorange', 'green', 'yellow']
f, ax1 = plt.subplots(1, figsize=(6, 6))
plt.subplots_adjust(left=None,
bottom=0.2,
right=None,
top=None,
wspace=None,
hspace=None)
#draw scatters
for X_transformed, title in [(X_pca, "PCA")]:
for color, i, group_name in zip(colors, range(len(group_names)),
group_names):
plt.scatter(X_transformed[y == i, 0],
X_transformed[y == i, 1],
color=color,
lw=2,
label=group_name)
plt.title(title + " of bacterium dataset")
plt.legend(loc="best", shadow=False, scatterpoints=1)
plt.axis([-1, 1, -0.5, 0.5])
global ann
ann = []
#set annotation
def annotation():
for label, x, y in zip(bacterium['sample'], X_pca[:, 0], X_pca[:, 1]):
ann.append(
ax1.annotate(label,
xy=(x, y),
xytext=(-2, 2),
textcoords='offset points',
ha='right',
va='bottom',
fontsize=6,
color='gray'))
global txt, arr
txt = []
arr = []
#loadings
def loading(load):
for i in range(len(pca.components_[0])):
x, y = pca.components_[0][i], pca.components_[1][i]
if x > load or y > load:
arr.append(
ax1.arrow(0,
0,
x * 0.5,
y * 0.5,
color='coral',
width=0.001,
head_width=0.01))
txt.append(
ax1.text(x * 0.25,
y * 0.25,
csv_list[i + 1][0],
color='coral',
ha='center',
va='center',
fontsize=7))
f.canvas.draw()
axloading = plt.axes([0.20, 0.10, 0.62, 0.03],
axisbg='lightgoldenrodyellow')
sloading = Slider(axloading,
'loading',
0.05,
0.5,
valinit=0.3,
color='coral')
def update(val):
for i in range(len(txt)):
txt[i].remove()
arr[i].remove()
arr[:] = []
txt[:] = []
load = sloading.val
loading(load)
f.canvas.draw_idle()
sloading.on_changed(update)
showax = plt.axes([0.7, 0.025, 0.1, 0.04])
hideax = plt.axes([0.8, 0.025, 0.1, 0.04])
bshow = Button(showax,
'show',
color='lightgoldenrodyellow',
hovercolor='0.975')
bhide = Button(hideax,
'hide',
color='lightgoldenrodyellow',
hovercolor='0.975')
def show(event):
annotation()
def hide(event):
for i in range(len(txt)):
txt[i].remove()
arr[i].remove()
arr[:] = []
txt[:] = []
for i in range(len(ann)):
ann[i].remove()
ann[:] = []
f.canvas.draw()
bshow.on_clicked(show)
bhide.on_clicked(hide)
plt.show()