def column_parser(lines):
"""Parser column format"""
record = False
result = []
struct = []
seq = ''
for line in lines:
if line.startswith('; ------'): #structure part beginns
record = True
continue
if line.startswith('; ******'): #structure part ends
record = False
struct = adjust_base(struct,-1)
struct = Pairs(struct).directed()#remove duplicates
struct.sort()
result.append([seq,struct])
struct = []
seq = ''
continue
if record:
sline = line.split()
if sline[4] == '.': #skip not paired
seq = ''.join([seq,sline[1]])
continue
seq = ''.join([seq,sline[1]])
pair = (int(sline[3]),int(sline[4])) #(alignpos,align_bp)
struct.append(pair)
return result
def adjust_pairs_from_mapping(pairs, mapping):
"""Returns new Pairs object with numbers adjusted according to map
pairs: list of tuples or Pairs object
mapping: dictionary containing mapping of positions from
one state to the other (e.g. ungapped to gapped)
For example:
{0: 0, 1: 1, 2: 3, 3: 4, 4: 6, 5: 7, 6: 9, 7: 10, 8: 12}
When the Pairs object corresponds to an ungapped sequence and
you want to insert gaps, use a mapping from ungapped to gapped.
When the Pairs object corresponds to a gapped sequence and you
want to degap it, use a mapping from gapped to ungapped.
"""
result = Pairs()
for x,y in pairs:
if x is None:
new_x = None
elif x not in mapping:
continue
else:
new_x = mapping[x]
if y is None:
new_y = None
elif y not in mapping:
continue
else:
new_y = mapping[y]
result.append((new_x, new_y))
return result
def adjust_pairs_from_mapping(pairs, mapping):
"""Returns new Pairs object with numbers adjusted according to map
pairs: list of tuples or Pairs object
mapping: dictionary containing mapping of positions from
one state to the other (e.g. ungapped to gapped)
For example:
{0: 0, 1: 1, 2: 3, 3: 4, 4: 6, 5: 7, 6: 9, 7: 10, 8: 12}
When the Pairs object corresponds to an ungapped sequence and
you want to insert gaps, use a mapping from ungapped to gapped.
When the Pairs object corresponds to a gapped sequence and you
want to degap it, use a mapping from gapped to ungapped.
"""
result = Pairs()
for x, y in pairs:
if x is None:
new_x = None
elif x not in mapping:
continue
else:
new_x = mapping[x]
if y is None:
new_y = None
elif y not in mapping:
continue
else:
new_y = mapping[y]
result.append((new_x, new_y))
return result
def adjust_base(pairs, offset):
"""Returns new Pairs with values shifted by offset
pairs: Pairs object or list of tuples
offset: integer
Adjusts the base of a pairs object or a list of pairs according to
the given offset.
There's no validation in here! It is possible negative values are
returned -> user responsibility.
This method treats all pairs as equal. It'll return a pairs object
of exactly the same length as the input, including pairs containing
None, and duplicates.
Example: adjust_base(Pairs([(2,8),(4,None)]), 2) --> [(4,10),(6,None)]
"""
if not isinstance(offset, int):
raise PairsAdjustmentError("adjust_base: offset should be integer")
result = Pairs()
for x, y in pairs:
if x is not None:
new_x = x + offset
else:
new_x = x
if y is not None:
new_y = y + offset
else:
new_y = y
result.append((new_x, new_y))
assert len(result) == len(pairs)
return result
def column_parser(lines):
"""Parser column format"""
record = False
result = []
struct = []
seq = ''
for line in lines:
if line.startswith('; ------'): #structure part beginns
record = True
continue
if line.startswith('; ******'): #structure part ends
record = False
struct = adjust_base(struct, -1)
struct = Pairs(struct).directed() #remove duplicates
struct.sort()
result.append([seq, struct])
struct = []
seq = ''
continue
if record:
sline = line.split()
if sline[4] == '.': #skip not paired
seq = ''.join([seq, sline[1]])
continue
seq = ''.join([seq, sline[1]])
pair = (int(sline[3]), int(sline[4])) #(alignpos,align_bp)
struct.append(pair)
return result
def adjust_base(pairs, offset):
"""Returns new Pairs with values shifted by offset
pairs: Pairs object or list of tuples
offset: integer
Adjusts the base of a pairs object or a list of pairs according to
the given offset.
There's no validation in here! It is possible negative values are
returned -> user responsibility.
This method treats all pairs as equal. It'll return a pairs object
of exactly the same length as the input, including pairs containing
None, and duplicates.
Example: adjust_base(Pairs([(2,8),(4,None)]), 2) --> [(4,10),(6,None)]
"""
if not isinstance(offset, int):
raise PairsAdjustmentError("adjust_base: offset should be integer")
result = Pairs()
for x, y in pairs:
if x is not None:
new_x = x + offset
else:
new_x = x
if y is not None:
new_y = y + offset
else:
new_y = y
result.append((new_x, new_y))
assert len(result) == len(pairs)
return result
def parse_residues(residue_lines, num_base, unpaired_symbol):
"""Return RnaSequence and Pairs object from residue lines.
residue_lines -- list of lines or anything that behaves like it.
Lines should contain:
residue_position, residue_identiy, residue_partner.
num_base -- int, basis of the residue numbering. In bpseq files from
the CRW website, the numbering starts at 1.
unpaired_symbol -- string, symbol in the 'partner' column that indicates
that a base is unpaired. In bpseq files from the CRW website, the
unpaired_symbol is '0'. This parameter should be a string to allow
other symbols that can't be casted to an integer to indicate
unpaired bases.
Checks for double entries both in the sequence and the structure, and
checks that the structre is valid in the sense that if (up,down) in there,
that (down,up) is the same.
"""
#create dictionary/list for sequence and structure
seq_dict = {}
pairs = Pairs()
for line in residue_lines:
try:
pos, res, partner = line.strip().split()
if partner == unpaired_symbol:
# adjust pos, not partner
pos = int(pos) - num_base
partner = None
else:
# adjust pos and partner
pos = int(pos) - num_base
partner = int(partner) - num_base
pairs.append((pos, partner))
#fill seq_dict
if pos in seq_dict:
raise BpseqParseError(\
"Double entry for residue %s (%s in bpseq file)"\
%(str(pos), str(pos+1)))
else:
seq_dict[pos] = res
except ValueError:
raise BpseqParseError("Failed to parse line: %s" % (line))
#check for conflicts, remove unpaired bases
if pairs.hasConflicts():
raise BpseqParseError("Conflicts in the list of basepairs")
pairs = pairs.directed()
pairs.sort()
# construct sequence from seq_dict
seq = RnaSequence(construct_sequence(seq_dict))
return seq, pairs
def parse_residues(residue_lines, num_base, unpaired_symbol):
"""Return RnaSequence and Pairs object from residue lines.
residue_lines -- list of lines or anything that behaves like it.
Lines should contain:
residue_position, residue_identiy, residue_partner.
num_base -- int, basis of the residue numbering. In bpseq files from
the CRW website, the numbering starts at 1.
unpaired_symbol -- string, symbol in the 'partner' column that indicates
that a base is unpaired. In bpseq files from the CRW website, the
unpaired_symbol is '0'. This parameter should be a string to allow
other symbols that can't be casted to an integer to indicate
unpaired bases.
Checks for double entries both in the sequence and the structure, and
checks that the structre is valid in the sense that if (up,down) in there,
that (down,up) is the same.
"""
#create dictionary/list for sequence and structure
seq_dict = {}
pairs = Pairs()
for line in residue_lines:
try:
pos, res, partner = line.strip().split()
if partner == unpaired_symbol:
# adjust pos, not partner
pos = int(pos) - num_base
partner = None
else:
# adjust pos and partner
pos = int(pos) - num_base
partner = int(partner) - num_base
pairs.append((pos,partner))
#fill seq_dict
if pos in seq_dict:
raise BpseqParseError(\
"Double entry for residue %s (%s in bpseq file)"\
%(str(pos), str(pos+1)))
else:
seq_dict[pos] = res
except ValueError:
raise BpseqParseError("Failed to parse line: %s"%(line))
#check for conflicts, remove unpaired bases
if pairs.hasConflicts():
raise BpseqParseError("Conflicts in the list of basepairs")
pairs = pairs.directed()
pairs.sort()
# construct sequence from seq_dict
seq = RnaSequence(construct_sequence(seq_dict))
return seq, pairs
def ct_parser(lines=None):
"""Ct format parser
Takes lines from a ct file as input
Returns a list containing sequence,structure and if available the energy.
[[seq1,[struct1],energy1],[seq2,[struct2],energy2],...]
"""
count = 0
length = ''
energy = None
seq = ''
struct = []
result = []
for line in lines:
count+=1
sline = line.split(None,6) #sline = split line
if count==1 or new_struct(line):#first line or new struct line.
if count > 1:
struct = adjust_base(struct,-1)
struct = Pairs(struct).directed()
struct.sort()
if energy is not None:
result.append([seq,struct,energy])
energy = None
else:
result.append([seq,pairs])
struct = []
seq = ''
#checks if energy for predicted struct is given
if sline.__contains__('dG') or sline.__contains__('ENERGY'):
energy = atof(sline[3])
if sline.__contains__('Structure'):
energy = atof(sline[2])
else:
seq = ''.join([seq,sline[1]])
if not int(sline[4]) == 0:#unpaired base
pair = ( int(sline[0]),int(sline[4]) )
struct.append(pair)
#structs are one(1) based, adjust to zero based
struct = adjust_base(struct,-1)
struct = Pairs(struct).directed()
struct.sort()
if energy is not None:
result.append([seq,struct,energy])
else:
result.append([seq,struct])
return result
def ct_parser(lines=None):
"""Ct format parser
Takes lines from a ct file as input
Returns a list containing sequence,structure and if available the energy.
[[seq1,[struct1],energy1],[seq2,[struct2],energy2],...]
"""
count = 0
length = ''
energy = None
seq = ''
struct = []
result = []
for line in lines:
count+=1
sline = line.split(None,6) #sline = split line
if count==1 or new_struct(line):#first line or new struct line.
if count > 1:
struct = adjust_base(struct,-1)
struct = Pairs(struct).directed()
struct.sort()
if energy is not None:
result.append([seq,struct,energy])
energy = None
else:
result.append([seq,pairs])
struct = []
seq = ''
#checks if energy for predicted struct is given
if sline.__contains__('dG') or sline.__contains__('ENERGY'):
energy = atof(sline[3])
if sline.__contains__('Structure'):
energy = atof(sline[2])
else:
seq = ''.join([seq,sline[1]])
if not int(sline[4]) == 0:#unpaired base
pair = ( int(sline[0]),int(sline[4]) )
struct.append(pair)
#structs are one(1) based, adjust to zero based
struct = adjust_base(struct,-1)
struct = Pairs(struct).directed()
struct.sort()
if energy is not None:
result.append([seq,struct,energy])
else:
result.append([seq,struct])
return result
def insert_gaps_in_pairs(pairs, gap_list):
"""Adjusts numbering in pairs according to the gap list.
pairs: Pairs object
gap_list: list of integers, gap positions in a sequence
The main assumptionis that all positions in pairs correspond to
ungapped positions. If this is not true, the result will be meaningless.
"""
if not gap_list:
new = Pairs()
new.extend(pairs)
return new
ungapped = []
for idx in range(max(gap_list)+2):
if idx not in gap_list:
ungapped.append(idx)
new = Pairs()
for x,y in pairs:
if x is not None:
try:
new_x = ungapped[x]
except IndexError:
new_x = ungapped[-1] + (x-len(ungapped)+1)
else:
new_x = x
if y is not None:
try:
new_y = ungapped[y]
except IndexError:
new_y = ungapped[-1] + (y-len(ungapped)+1)
else:
new_y = y
new.append((new_x, new_y))
return new
def delete_gaps_from_pairs(pairs, gap_list):
"""Returns Pairs object with pairs adjusted to gap_list
pairs: list of tuples or Pairs object
gap_list: list or array of gapped positions that should be removed
from the pairs object
Base pairs of which one of the partners or both of them are in
the gap list are removed. If both of them are not in the gap_list, the
numbering is adjusted according to the gap_list.
When at least one of the two pair members is in the gap_list, the
pair will be removed. The rest of the structure will be left
intact. Pairs containing None, duplicates, pseudoknots, and
conflicts will be maintained and adjusted according to the gap_list.
"""
if not gap_list:
result = Pairs()
result.extend(pairs)
return result
g = array(gap_list)
result = Pairs()
for up, down in pairs:
if up in g or down in g:
continue
else:
if up is not None:
new_up = up - g.searchsorted(up)
else:
new_up = up
if down is not None:
new_down = down - g.searchsorted(down)
else:
new_down = down
result.append((new_up, new_down))
return result
def insert_gaps_in_pairs(pairs, gap_list):
"""Adjusts numbering in pairs according to the gap list.
pairs: Pairs object
gap_list: list of integers, gap positions in a sequence
The main assumptionis that all positions in pairs correspond to
ungapped positions. If this is not true, the result will be meaningless.
"""
if not gap_list:
new = Pairs()
new.extend(pairs)
return new
ungapped = []
for idx in range(max(gap_list) + 2):
if idx not in gap_list:
ungapped.append(idx)
new = Pairs()
for x, y in pairs:
if x is not None:
try:
new_x = ungapped[x]
except IndexError:
new_x = ungapped[-1] + (x - len(ungapped) + 1)
else:
new_x = x
if y is not None:
try:
new_y = ungapped[y]
except IndexError:
new_y = ungapped[-1] + (y - len(ungapped) + 1)
else:
new_y = y
new.append((new_x, new_y))
return new
def delete_gaps_from_pairs(pairs, gap_list):
"""Returns Pairs object with pairs adjusted to gap_list
pairs: list of tuples or Pairs object
gap_list: list or array of gapped positions that should be removed
from the pairs object
Base pairs of which one of the partners or both of them are in
the gap list are removed. If both of them are not in the gap_list, the
numbering is adjusted according to the gap_list.
When at least one of the two pair members is in the gap_list, the
pair will be removed. The rest of the structure will be left
intact. Pairs containing None, duplicates, pseudoknots, and
conflicts will be maintained and adjusted according to the gap_list.
"""
if not gap_list:
result = Pairs()
result.extend(pairs)
return result
g = array(gap_list)
result = Pairs()
for up, down in pairs:
if equal(g, up).any() or equal(g, down).any():
continue
if up is not None:
new_up = up - g.searchsorted(up)
else:
new_up = up
if down is not None:
new_down = down - g.searchsorted(down)
else:
new_down = down
result.append((new_up, new_down))
return result
