allpy: lib/block.py annotate

allpy

annotate lib/block.py @ 143:004c2f6c45ac

project.from_fasta improvements

author	boris <bnagaev@gmail.com>
date	Sun, 24 Oct 2010 19:05:31 +0400
parents	0b366ce3257f
children	8f1d8ece31af

rev	line source
BurkovBA@0	1 #!usr/bin/python
BurkovBA@0	2
BurkovBA@0	3 import sys
BurkovBA@0	4
BurkovBA@0	5 import project
BurkovBA@0	6 import sequence
BurkovBA@0	7 import monomer
bnagaev@116	8 import config
bnagaev@121	9 from graph import Graph
BurkovBA@0	10
BurkovBA@0	11 class Block(object):
BurkovBA@0	12 """
BurkovBA@0	13 Mandatory data:
BurkovBA@0	14 * self.project -- project object, which the block belongs to
BurkovBA@1	15 * self.sequences - set of sequence objects that contain monomers
BurkovBA@0	16 and/or gaps, that constitute the block
bnagaev@115	17 * self.positions -- sorted list of positions of the project.alignment that
BurkovBA@1	18 are included in the block
bnagaev@116	19
bnagaev@132	20 Don't change self.sequences -- it may be a link to other block.sequences
bnagaev@132	21
BurkovBA@0	22 How to create a new block:
BurkovBA@0	23 >>> import project
BurkovBA@0	24 >>> import block
BurkovBA@0	25 >>> proj = project.Project(open("test.fasta"))
bnagaev@114	26 >>> block1 = block.Block(proj)
BurkovBA@0	27 """
BurkovBA@0	28
bnagaev@112	29 def __init__(self, project, sequences=None, positions=None):
bnagaev@112	30 """
bnagaev@112	31 Builds new block from project
bnagaev@112	32 if sequences==None, all sequences are used
bnagaev@112	33 if positions==None, all positions are used
bnagaev@112	34 """
bnagaev@112	35 if sequences == None:
bnagaev@132	36 sequences = set(project.sequences) # copy
bnagaev@112	37 if positions == None:
bnagaev@112	38 positions = range(len(project))
BurkovBA@73	39 self.project = project
BurkovBA@73	40 self.sequences = sequences
BurkovBA@73	41 self.positions = positions
BurkovBA@0	42
bnagaev@137	43 def save_fasta(self, out_file, long_line=60, gap='-'):
bnagaev@112	44 """
bnagaev@112	45 Saves alignment to given file in fasta-format
bnagaev@112	46 Splits long lines to substrings of length=long_line
bnagaev@112	47 To prevent this, set long_line=None
bnagaev@112	48
BurkovBA@0	49 No changes in the names, descriptions or order of the sequences
BurkovBA@0	50 are made.
BurkovBA@0	51 """
BurkovBA@0	52 for sequence in self.sequences:
bnagaev@112	53 out_file.write(">%(name)s %(description)s \n" % sequence.__dict__)
bnagaev@113	54 alignment_monomers = self.project.alignment[sequence]
bnagaev@115	55 block_monomers = [alignment_monomers[i] for i in self.positions]
bnagaev@113	56 string = ''.join([m.type.code1 if m else '-' for m in block_monomers])
bnagaev@112	57 if long_line:
bnagaev@112	58 for i in range(0, len(string) // long_line + 1):
bnagaev@112	59 out_file.write("%s \n" % string[ilong_line : ilong_line + long_line])
bnagaev@112	60 else:
bnagaev@112	61 out_file.write("%s \n" % string)
bnagaev@116	62
bnagaev@127	63 def geometrical_cores(self, max_delta=config.delta,
bnagaev@129	64 timeout=config.timeout, minsize=config.minsize,
bnagaev@129	65 ac_new_atoms=config.ac_new_atoms,
bnagaev@129	66 ac_count=config.ac_count):
bnagaev@116	67 """
bnagaev@129	68 returns length-sorted list of blocks, representing GCs
bnagaev@126	69
bnagaev@129	70 max_delta -- threshold of distance spreading
bnagaev@129	71 timeout -- Bron-Kerbosh timeout (then fast O(n ln n) algorithm)
bnagaev@129	72 minsize -- min size of each core
bnagaev@129	73 ac_new_atoms -- min part or new atoms in new alternative core
bnagaev@129	74 current GC is compared with each of already selected GCs
bnagaev@129	75 if difference is less then ac_new_atoms, current GC is skipped
bnagaev@129	76 difference = part of new atoms in current core
bnagaev@129	77 ac_count -- max number of cores (including main core)
bnagaev@130	78 -1 means infinity
bnagaev@120	79 If more than one pdb chain for some sequence provided, consider all of them
bnagaev@120	80 cost is calculated as 1 / (delta + 1)
bnagaev@120	81 delta in [0, +inf) => cost in (0, 1]
bnagaev@116	82 """
bnagaev@117	83 nodes = self.positions
bnagaev@117	84 lines = {}
bnagaev@116	85 for i in self.positions:
bnagaev@116	86 for j in self.positions:
bnagaev@117	87 if i < j:
bnagaev@120	88 distances = []
bnagaev@120	89 for sequence in self.sequences:
bnagaev@120	90 for chain in sequence.pdb_chains:
bnagaev@120	91 m1 = self.project.alignment[sequence][i]
bnagaev@120	92 m2 = self.project.alignment[sequence][j]
bnagaev@122	93 if m1 and m2:
bnagaev@122	94 ca1 = m1.pdb_residues[chain]['CA']
bnagaev@122	95 ca2 = m2.pdb_residues[chain]['CA']
bnagaev@122	96 d = ca1 - ca2 # Bio.PDB feature
bnagaev@122	97 distances.append(d)
bnagaev@122	98 if len(distances) >= 2:
bnagaev@122	99 delta = max(distances) - min(distances)
bnagaev@122	100 if delta <= max_delta:
bnagaev@122	101 lines[Graph.line(i, j)] = 1.0 / (1.0 + max_delta)
bnagaev@120	102 graph = Graph(nodes, lines)
bnagaev@129	103 cliques = graph.cliques(timeout=timeout, minsize=minsize)
bnagaev@129	104 GCs = []
bnagaev@129	105 for clique in cliques:
bnagaev@129	106 for GC in GCs:
bnagaev@129	107 if len(clique - set(GC.positions)) < ac_new_atoms * len(clique):
bnagaev@129	108 break
bnagaev@129	109 else:
bnagaev@132	110 GCs.append(Block(self.project, self.sequences, clique))
bnagaev@130	111 if ac_count != -1 and len(GCs) >= ac_count:
bnagaev@129	112 break
bnagaev@129	113 return GCs
bnagaev@123	114
bnagaev@137	115 def xstring(self, x='X', gap='-'):
bnagaev@123	116 """
bnagaev@137	117 Returns string consisting of gap chars and chars x at self.positions
bnagaev@123	118 Length of returning string = length of project
bnagaev@123	119 """
bnagaev@123	120 monomers = [False] * len(self.project)
bnagaev@123	121 for i in self.positions:
bnagaev@123	122 monomers[i] = True
bnagaev@137	123 return ''.join([x if m else gap for m in monomers])
bnagaev@134	124
bnagaev@137	125 def save_xstring(self, out_file, name, description='', x='X', gap='-'):
bnagaev@134	126 """
bnagaev@134	127 Save xstring and name in fasta format
bnagaev@134	128 """
bnagaev@134	129 out_file.write(">%(name)s %(description)s \n" % \
bnagaev@134	130 {'name':name, 'description':description})
bnagaev@134	131
bnagaev@142	132 out_file.write("%(xstring)s \n" % {'xstring':self.xstring(x=x, gap=gap)})