In [1]:
%matplotlib nbagg
import matplotlib.pyplot as plt
In [2]:
%cd ../data/

/home/explorer/BMS270/data
In [3]:
from csv import reader, excel_tab
In [4]:
orfs = []
names = []
data = []
fin = reader(open("GSE88801_kallisto_TPMs_thresh10.cdt"),dialect=excel_tab)
header = next(fin)[2:]
for row in fin:
    orfs.append(row[0])
    names.append(row[1])
    data.append([float(i) for i in row[2:]])
In [5]:
import numpy as np
In [6]:
A = np.array(data)
In [7]:
A.shape, A.dtype
Out[7]:
((9939, 36), dtype('float64'))
In [8]:
np.mean(A)
Out[8]:
4.7600470545758755
In [9]:
m = np.mean(A,axis=0)
In [10]:
m.shape
Out[10]:
(36,)
In [11]:
m = np.mean(A,axis=1)
In [12]:
m.shape
Out[12]:
(9939,)
In [13]:
m = m.reshape((9939,1))
In [14]:
m.shape
Out[14]:
(9939, 1)
In [15]:
D = A - m
In [16]:
D.shape
Out[16]:
(9939, 36)
In [18]:
fig = plt.figure()
plt.imshow(A, interpolation="none", aspect="auto")
plt.colorbar()
Out[18]:
<matplotlib.colorbar.Colorbar at 0x7fa7b2ba5cf8>
In [19]:
fig = plt.figure()
plt.imshow(D, interpolation="none", aspect="auto")
plt.colorbar()
Out[19]:
<matplotlib.colorbar.Colorbar at 0x7fa7b11ef860>
In [20]:
from numpy.linalg import norm
In [21]:
s = norm(D, axis=1)
In [22]:
s.shape
Out[22]:
(9939,)
In [23]:
s = s.reshape((9939,1))
In [24]:
C = D/s
In [26]:
fig = plt.figure()
plt.imshow(C, interpolation="none", aspect="auto")
plt.colorbar()
Out[26]:
<matplotlib.colorbar.Colorbar at 0x7fa7b10a2c88>
In [28]:
header[:10]
Out[28]:
['BMDM_Live_1_4h',
 'BMDM_Live_1_24h',
 'BMDM_Live_2_4h',
 'BMDM_Live_2_24h',
 'BMDM_Live_3_4h',
 'BMDM_Live_3_24h',
 'BMDM_Dead_1_4h',
 'BMDM_Dead_1_24h',
 'BMDM_Dead_2_4h',
 'BMDM_Dead_2_24h']
In [29]:
fig = plt.figure()
plt.plot(A[:100,0],A[:100,2],"bo")
Out[29]:
[<matplotlib.lines.Line2D at 0x7fa7b0f34048>]
In [30]:
fig = plt.figure()
plt.plot(D[:100,0],D[:100,2],"bo")
Out[30]:
[<matplotlib.lines.Line2D at 0x7fa7b0f20d30>]
In [31]:
fig = plt.figure()
plt.plot(C[:100,0],C[:100,2],"bo")
Out[31]:
[<matplotlib.lines.Line2D at 0x7fa7b0e8e0f0>]
In [32]:
corr = np.dot(C,C.T)
In [33]:
corr.shape
Out[33]:
(9939, 9939)
In [35]:
fig = plt.figure()
plt.imshow(corr[:1000,:1000],interpolation="none",aspect="auto")
plt.colorbar()
Out[35]:
<matplotlib.colorbar.Colorbar at 0x7fa7b0d2b5c0>
In [36]:
np.pi
Out[36]:
3.141592653589793
In [37]:
theta = np.pi/2
In [38]:
rotation = np.array([[np.cos(theta),-np.sin(theta)],
                     [np.sin(theta),np.cos(theta)]])
In [39]:
rotation
Out[39]:
array([[  6.12323400e-17,  -1.00000000e+00],
       [  1.00000000e+00,   6.12323400e-17]])
In [40]:
test_points = np.array([[1,2,3,4,5],
                       [-1,-2,3,8,7]], dtype="float")
In [41]:
test_points
Out[41]:
array([[ 1.,  2.,  3.,  4.,  5.],
       [-1., -2.,  3.,  8.,  7.]])
In [43]:
fig = plt.figure()
plt.plot(test_points[0],test_points[1],"bo")
rotated = np.dot(rotation,test_points)
plt.plot(rotated[0],rotated[1],"ro")
Out[43]:
[<matplotlib.lines.Line2D at 0x7fa7b0c3aeb8>]
In [44]:
import Bio.Cluster as Pycluster
In [45]:
Dsub = D[:1000,:]
Dsub.shape
Out[45]:
(1000, 36)
In [46]:
%%time
tree = Pycluster.treecluster(Dsub, dist="u")

CPU times: user 252 ms, sys: 8 ms, total: 260 ms
Wall time: 263 ms
In [47]:
Dsub = D[:2000,:]
Dsub.shape
Out[47]:
(2000, 36)
In [48]:
%%time
tree = Pycluster.treecluster(Dsub, dist="u")

CPU times: user 1.7 s, sys: 8 ms, total: 1.7 s
Wall time: 1.74 s
In [49]:
%%time
tree = Pycluster.treecluster(D, dist="u")

CPU times: user 3min 33s, sys: 1.7 s, total: 3min 34s
Wall time: 3min 40s
In [50]:
record = Pycluster.Record()
record.data = D
record.geneid = orfs[:]
record.genename = names[:]
record.gweight = None
record.gorder = None
record.expid = header[:]
record.eweight = None
record.eorder = None
record.uniqid = "UNIQID"
record.save("GSE88801_kallisto_TPMs_thresh10.centered.um", geneclusters = tree)

rsync -av -e 'ssh -p 8022' explorer@localhost:BMS270/data/"\*.centered.um.\*" ./
In [ ]: