def draw_swarm(self,columns,path=None): values = [] for column in self[columns]: values.append(self[column].values) fig = pyplot.figure() ax = fig.add_subplot(111) beeswarm(values, method="square", labels=columns, ax=ax) if path is not None: fig.savefig(path) else: return fig
def Beeswarm(self):
colors = ['red','cyan','green','magenta','blue','black']
# boardcasting color cycle
colors = self.data.num/len(colors)*colors+colors[0:self.data.num%len(colors)]
# Data
datavect = []
[datavect.append(self.data.series['y'+str(n+1)]) for n in range(self.data.num)]
print(datavect)
# names
datanames = []
[datanames.append(self.data.names['y'+str(n+1)]) for n in range(self.data.num)]
self.bs, ax = beeswarm(datavect, method=self.Style,
labels=datanames,
col=colors)
# Format style
# make sure axis tickmark points out
ax.tick_params(axis='both',direction='out')
ax.spines['right'].set_visible(False)
ax.spines['top'].set_visible(False)
ax.xaxis.set_ticks_position('bottom')
ax.yaxis.set_ticks_position('left')
# save current figure handle
self.fig = plt.gcf()
self.axs = ax
# Do annotation: compare significance
X = ax.get_xticks()
Y = [max(x) for x in datavect]
self.label_diff(0,1,'p=0.0370',X,Y)
self.label_diff(1,2,'p<0.0001',X,Y)
def plot(self):
bs, ax = beeswarm(
self.data,
method="center",
s=.5,
__s=10,
labelrotation="horizontal",
ax=self.figure.add_subplot(*self.plot_tuple), labels=self.x_labels,
col="white",
edgecolors="black"
)
if self.plot_tuple[2] != 1:
ax.set_xlabel("Day & frequency")
ax.set_ylabel("Colony Size (cells)")
ax.set_title(self.graph_label)
ax.set_ylim(bottom=0, top=self.cap)
ax.set_yticks((0, 10, 20, 30, 40))
ax.grid(color='black', linestyle='-', linewidth=.5, axis="y")
def analyze_transcript_position(feature_counts, feature_type, cutoff=0):
feature_counts_indexed = feature_counts.set_index(feature_counts[("Transcript0","Transcript0")])
intron_count = 1
read_dict = {}
#Pick every row with intron number
while intron_count < 10:
read_list = []
for row in feature_counts_indexed.iterrows():
if row[1][1] == intron_count:
if feature_type == "Exon":
read_list.append(row[1][2])
elif feature_type == "Intron":
read_list.append(row[1][3])
else: print "Unknown feature type"
#Append to dictionary - keys are intron number (or exon) and values are list of read counts
print feature_type+" position "+str(intron_count)
print "Before cutoff:"+str(len(read_list))+" "+feature_type+"s analyzed"
read_list = [math.log(x, 10) for x in read_list if x > cutoff]
read_dict[intron_count] = read_list
print "After cutoff:"+str(len(read_list))+" "+feature_type+"s analyzed \n"
intron_count += 1
#Extract keys and lists for use in beeswarm
values_list = []
name_list = []
for intron_number, read_list in read_dict.iteritems():
name_list.append(intron_number)
values_list.append(read_list)
color_list = ["red", "orange", "yellow", "green", "teal", "blue", "purple", "indigo", "grey"]
#Make beeswarm - x-axis is intron number and y-axis is counts
bs, ax = beeswarm(values_list, method="swarm", labels=name_list, col=color_list)
#Do I need to normalize to total RNA?
def beeswarm_plot(DataFrame_list, SampleTuple_list, DataFrame2=None, base=10, color_list="blue", select_random=False):
print SampleTuple_list
print len(SampleTuple_list)
values_list = []
name_list = []
median_list = []
a=0
while a < len(DataFrame_list):
n=0
while n < len(SampleTuple_list):
print SampleTuple_list[n]
values = get_ratios(DataFrame_list[a], SampleTuple_list[n][0], SampleTuple_list[n][1], log=True, base=base)
median_list.append(np.median(np.array(values)))
if select_random != False:
values = random.sample(values, select_random)
values_list.append(values)
name_list.append(SampleTuple_list[n][0])
n += 1
a += 1
print median_list
bs, ax = beeswarm(values_list, method="swarm", labels=name_list, col=color_list)
def Beeswarm(self):
colors = ['red', 'cyan', 'green', 'magenta', 'blue', 'black']
# boardcasting color cycle
colors = self.data.num / len(
colors) * colors + colors[0:self.data.num % len(colors)]
# Data
datavect = []
[
datavect.append(self.data.series['y' + str(n + 1)])
for n in range(self.data.num)
]
print(datavect)
# names
datanames = []
[
datanames.append(self.data.names['y' + str(n + 1)])
for n in range(self.data.num)
]
self.bs, ax = beeswarm(datavect,
method=self.Style,
labels=datanames,
col=colors)
# Format style
# make sure axis tickmark points out
ax.tick_params(axis='both', direction='out')
ax.spines['right'].set_visible(False)
ax.spines['top'].set_visible(False)
ax.xaxis.set_ticks_position('bottom')
ax.yaxis.set_ticks_position('left')
# save current figure handle
self.fig = plt.gcf()
self.axs = ax
# Do annotation: compare significance
X = ax.get_xticks()
Y = [max(x) for x in datavect]
self.label_diff(0, 1, 'p=0.0370', X, Y)
self.label_diff(1, 2, 'p<0.0001', X, Y)
def analyze_transcript_position(feature_counts, feature_type, cutoff=0):
feature_counts_indexed = feature_counts.set_index(feature_counts[("Transcript0","Transcript0")])
intron_count = 1
read_dict = {}
#Pick every row with intron number
while intron_count < 10:
read_list = []
for row in feature_counts_indexed.iterrows():
if row[1][1] == intron_count:
if feature_type == "Exon":
read_list.append(row[1][2])
elif feature_type == "Intron":
read_list.append(row[1][3])
else: print "Unknown feature type"
#Append to dictionary - keys are intron number (or exon) and values are list of read counts
print feature_type+" position "+str(intron_count)
print "Before cutoff:"+str(len(read_list))+" "+feature_type+"s analyzed"
read_list = [math.log(x, 10) for x in read_list if x > cutoff]
read_dict[intron_count] = read_list
print "After cutoff:"+str(len(read_list))+" "+feature_type+"s analyzed \n"
intron_count += 1
#Extract keys and lists for use in beeswarm
values_list = []
name_list = []
for intron_number, read_list in read_dict.iteritems():
name_list.append(intron_number)
values_list.append(read_list)
color_list = ["red", "orange", "yellow", "green", "teal", "blue", "purple", "indigo", "grey"]
#Make beeswarm - x-axis is intron number and y-axis is counts
bs, ax = beeswarm(values_list, method="swarm", labels=name_list, col=color_list)
#!/usr/bin/python
import numpy as np
import matplotlib.pyplot as plt
from beeswarm import *
import sys, getopt
from Bio import SeqIO
from Bio.Seq import Seq
from Bio.Alphabet import IUPAC
from collections import Counter
d1 = np.random.uniform(low=-3, high=3, size=100)
d2 = np.random.normal(size=100)
bs, ax = beeswarm([d1,d2], method="swarm", labels=["sample 1", "sample 2"], col=["blue","red"])
ax2 = plt.subplot(412)
ax2.set_title('Medium GC')
ax3 = plt.subplot(413)
ax3.set_title('High GC')
ax4 = plt.subplot(414)
ax4.set_title('High GC')
axes = [ax1, ax2, ax3, ax4]
beeswarm([Low100pg,Low10pg,Low1pg,Low100fg], method="swarm", labels=["100pg","10pg","1pg","100fg"], col=["black","red","green","blue"], ax=ax1)
beeswarm([Medium100pg,Medium10pg,Medium1pg,Medium100fg], method="swarm", labels=["100pg","10pg","1pg","100fg"], col=["black","red","green","blue"], ax=ax2)
beeswarm([High100pg,High10pg,High1pg,High100fg], method="swarm", labels=["100pg","10pg","1pg","100fg"], col=["black","red","green","blue"], ax=ax3)
beeswarm([Fake100pg,Fake10pg,Fake1pg,Fake100fg], method="swarm", labels=["100pg","10pg","1pg","100fg"], col=["black","red","green","blue"], ax=ax4)
# set y-axes
#ax2.set_xlim([0, 100])
ax1.set_ylim([-10, 6])
ax2.set_ylim([-10, 6])
ax3.set_ylim([-10, 6])
#ax4.set_ylim([-1.4, 0.4])
f, ax = plt.subplots(figsize=(12, 9))
sns.heatmap(corrmap, square=True)
'''
audio only, PHQ8_Binary
'''
pd_audio = pd.read_csv('/Users/mac/Downloads/avec2017/audio_fea_train_binary.csv', \
index_col=0)
pd_audio.shape
no_important = [867, 852, 795, 307, 196, 195, 206, 362, 419, 1009, 1192]
pd_important = pd_audio.ix[:, no_important]
no = 9
bs, ax = beeswarm([pd_important[pd_important.y==0].iloc[:,no].values, \
pd_important[pd_important.y==1].iloc[:,no].values], \
method = 'swarm', labels = ['0', '1'], col = ['blue', 'red'])
'''
text only, PHQ8_Score
'''
pd_text = pd.read_csv('/Users/mac/Downloads/avec2017/text_fea_train.csv', \
index_col=0)
pd_text.shape
no_important = [0, 1, 2, 3, 4, 5, 6, 7, 8, 10]
pd_important = pd_text.ix[:, no_important]
corrmap = pd_important.corr()
f, ax = plt.subplots(figsize=(12, 9))
sns.heatmap(corrmap, square=True)