In [1]:

    

# Fetch an example MSA from the Pfam database:
!wget https://pfam.xfam.org/family/PF06201/alignment/full/gzipped -O MSA.gz


    

    




--2018-07-23 23:39:35--  https://pfam.xfam.org/family/PF06201/alignment/full/gzipped
Resolving pfam.xfam.org (pfam.xfam.org)... 193.62.193.83
Connecting to pfam.xfam.org (pfam.xfam.org)|193.62.193.83|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 221647 (216K) [application/x-gzip]
Saving to: ‘MSA.gz’

MSA.gz              100%[===================>] 216,45K   210KB/s    in 1,0s    

2018-07-23 23:39:37 (210 KB/s) - ‘MSA.gz’ saved [221647/221647]

In [2]:

    

# Parse and filter sequence records into a SequenceData object:
import skmsa
sd = skmsa.read('MSA.gz').filtered() # use defaults (check docstrings for read and filter options)


    

In [3]:

    

#Convert sequence data to the most appropriate object, for example:
#a pandas dataframe:
df = sd.as_dataframe()
df.head()


    

    
Out[3]:







  

    

      

      
header
      
1
      
2
      
3
      
4
      
5
      
6
      
7
      
8
      
9
      
...
      
141
      
142
      
143
      
144
      
145
      
146
      
147
      
148
      
149
      
150
    
  
  

    

      
1
      
C5DGB7_LACTC/26-197
      
F
      
S
      
F
      
V
      
N
      
T
      
P
      
N
      
V
      
...
      
T
      
R
      
C
      
F
      
Y
      
L
      
E
      
M
      
R
      
G
    
    

      
2
      
A0A0D8Y9P7_DICVI/124-264
      
-
      
-
      
L
      
I
      
D
      
K
      
K
      
Q
      
M
      
...
      
T
      
E
      
I
      
E
      
K
      
L
      
K
      
I
      
F
      
G
    
    

      
5
      
A0A095CE52_CRYGR/34-192
      
W
      
S
      
H
      
I
      
D
      
R
      
D
      
N
      
V
      
...
      
S
      
Q
      
V
      
Y
      
F
      
I
      
G
      
L
      
K
      
G
    
    

      
7
      
H2ZHR8_CIOSA/125-269
      
-
      
-
      
-
      
-
      
-
      
-
      
-
      
-
      
-
      
...
      
T
      
K
      
I
      
Q
      
R
      
I
      
I
      
L
      
Y
      
G
    
    

      
8
      
K1QGE0_CRAGI/33-175
      
Y
      
T
      
K
      
I
      
N
      
M
      
E
      
A
      
V
      
...
      
T
      
K
      
I
      
Y
      
Y
      
I
      
G
      
L
      
K
      
G
    
  

5 rows × 128 columns

In [4]:

    

df.describe()


    

    
Out[4]:







  

    

      

      
header
      
1
      
2
      
3
      
4
      
5
      
6
      
7
      
8
      
9
      
...
      
141
      
142
      
143
      
144
      
145
      
146
      
147
      
148
      
149
      
150
    
  
  

    

      
count
      
854
      
854
      
854
      
854
      
854
      
854
      
854
      
854
      
854
      
854
      
...
      
854
      
854
      
854
      
854
      
854
      
854
      
854
      
854
      
854
      
854
    
    

      
unique
      
854
      
18
      
18
      
19
      
9
      
19
      
21
      
20
      
19
      
12
      
...
      
11
      
20
      
9
      
20
      
17
      
8
      
18
      
8
      
17
      
2
    
    

      
top
      
A0A0C3NG11_PHLGI/9-157
      
Y
      
S
      
Q
      
I
      
D
      
F
      
D
      
G
      
V
      
...
      
T
      
R
      
I
      
D
      
Y
      
I
      
G
      
L
      
K
      
G
    
    

      
freq
      
1
      
281
      
148
      
166
      
580
      
565
      
148
      
187
      
169
      
321
      
...
      
570
      
458
      
491
      
207
      
409
      
434
      
414
      
343
      
234
      
845
    
  

4 rows × 128 columns

In [5]:

    

# a NumPy ndarray:
X = sd.as_array()
X


    

    
Out[5]:





array([['F', 'S', 'F', ..., 'M', 'R', 'G'],
       ['-', '-', 'L', ..., 'I', 'F', 'G'],
       ['W', 'S', 'H', ..., 'L', 'K', 'G'],
       ...,
       ['Y', 'G', 'T', ..., 'F', 'K', 'G'],
       ['T', 'S', 'Q', ..., 'I', 'I', 'G'],
       ['Y', 'T', 'F', ..., 'F', 'R', 'G']], dtype='<U1')

In [6]:

    

# a numerically encoded array:
X = sd.as_encoded() # default encoding: map residue symbols to their index in the alignment alphabet
print(sd.alphabet)
X


    

    




-ACDEFGHIKLMNPQRSTVWY






    
Out[6]:





array([[ 5, 16,  5, ..., 11, 15,  6],
       [ 0,  0, 10, ...,  8,  5,  6],
       [19, 16,  7, ..., 10,  9,  6],
       ...,
       [20,  6, 17, ...,  5,  9,  6],
       [17, 16, 14, ...,  8,  8,  6],
       [20, 17,  5, ...,  5, 15,  6]])

In [7]:

    

# a one-hot encoded array (as a SciPy sparse matrix in CSR format):
X = sd.as_one_hot(sparse=True)
X


    

    
Out[7]:





<854x2540 sparse matrix of type '<class 'numpy.float64'>'
	with 107584 stored elements in Compressed Sparse Row format>