root/galaxy-central/eggs/bx_python-0.5.0_dev_f74aec067563-py2.6-macosx-10.6-universal-ucs2.egg/bx/align/sitemask/cpg.py

リビジョン 3, 2.8 KB (コミッタ: kohda, 14 年 前)

Install Unix tools  http://hannonlab.cshl.edu/galaxy_unix_tools/galaxy.html

行番号 
1"""
2Support for masking potential CpG sites in *pairwise* alignments.
3"""
4
5from _cpg import *
6from bx.align.sitemask import Masker
7from bx.filter import *
8import string
9
10# Restricted.  Only mask out sites that are defitely CpG
11class Restricted( Masker ):
12    def __init__( self, mask = '?' ):
13        self.mask = mask
14        self.masked = 0
15        self.total = 0
16       
17    def __call__( self, block ):
18        if not block: return block
19        if len(block.components) < 2:
20            return
21        cpglist = list_cpg_restricted( \
22            string.upper(block.components[0].text), \
23            string.upper(block.components[1].text) )
24
25        # now we have a fast list of CpG columns, iterate/mask
26        self.masked += len(cpglist)
27        self.total += len(block.components[0].text)
28        for component in block.components:
29            component.text = mask_columns( cpglist, component.text, self.mask )
30           
31        return block
32   
33# Inclusive. Mask out all sites that are not non-CpG sites.
34class Inclusive( Masker ):
35    def __init__( self, mask = '?' ):
36        self.mask = mask
37        self.masked = 0
38        self.total = 0
39       
40    def __call__( self, block ):
41        if not block: return block
42        if len(block.components) < 2:
43            return
44        cpglist = list_cpg( \
45            string.upper(block.components[0].text), \
46            string.upper(block.components[1].text) )
47       
48        self.masked += len( cpglist )
49        self.total += len( block.components[0].text )
50        for component in block.components:
51            component.text = mask_columns( cpglist, component.text, self.mask)
52           
53        return block
54
55#Mak nonCpG sites
56class nonCpG( Masker ):
57    def __init__( self, mask = '?' ):
58        self.mask = mask
59        self.masked = 0
60        self.total = 0
61       
62    def __call__( self, block ):
63        if not block: return block
64        if len(block.components) < 2:
65            return
66        noncpglist = list_non_cpg( \
67            string.upper(block.components[0].text), \
68            string.upper(block.components[1].text) )
69
70        # now we have a fast list of non-CpG columns, iterate/mask
71        self.masked += len(noncpglist)
72        self.total += len(block.components[0].text)
73        for component in block.components:
74            component.text = mask_columns( noncpglist, component.text, self.mask )
75           
76        return block
77   
78def mask_columns( masklist, text, mask ):
79    templist = list()
80    for position in masklist:
81        if text[position] != "-":
82            templist.append(position)
83    templist.append(len(text)) # Add the end of the text
84    #cut string
85    newtext = list()
86    c = 0
87    for position in templist:
88        newtext.append(text[c:position])
89        c = position + 1 # Gaps have len = 1
90    joinedtext = mask.join(newtext)
91    return joinedtext
Note: リポジトリブラウザについてのヘルプは TracBrowser を参照してください。