def execute(self, input_commstructs):
if input_commstructs == None or len(input_commstructs) == 0:
raise ExecutionException("BlockProcessor.execute : No inputs")
input_commstruct = input_commstructs[0]
# retrieve the processor parameters
self.windowSize = self.getParameterAsint(
BlockProcessor.WINDOW_SIZE_PARAM)
self.residuConservationLimit = self.getParameterAsfloat(
BlockProcessor.RESIDU_CONSERVATION_LIMIT_PARAM)
self.windowConservationLimit = self.getParameterAsfloat(
BlockProcessor.WINDOW_CONSERVATION_LIMIT_PARAM)
algo = self.getParameter(BlockProcessor.ALGORITHM_PARAM, False)
if algo != None:
self.algorithm = algo.lower()
referenceSpecies = self.getParameter(
BlockProcessor.REFERENCE_SPECIES_PARAM)
desired_species_line = self.getParameter(
BlockProcessor.DESIRED_SPECIES_LIST_PARAM, False)
Log.trace("BlockProcessor.execute : Chosen Algorithm is '" +
self.algorithm + "'")
self.desiredSpeciesList = []
self.desiredSpeciesList.append(referenceSpecies)
if desired_species_line != None:
self.desiredSpeciesList.extend(desired_species_line.split())
# Analyze the conserved region in each MSA
# If 'None' algorithm is chosen, the entire MSA is considered as conserved
for bed_seq in input_commstruct.bedToMA.keys():
for alignment in input_commstruct.bedToMA[bed_seq]:
pwm = PWM()
pwm.initFromAlignment(alignment, self.desiredSpeciesList)
if self.algorithm != BlockProcessor.ALGORITHM_NONE_VALUE:
self.analyzeConservedBlocks(pwm, alignment)
else:
new_block = Motif(0, alignment.totalLength, "", pwm)
new_block.composeName(alignment.name)
alignment.addMotif(new_block, True)
return input_commstruct
def getMotifDefinitions(self, motif_name_list, database_file_path):
motif_list = []
for name in motif_name_list:
motif_list.append(Motif(0, 0, name, None))
MotifUtils.getMotifsPWMFromJasparTF(motif_list, database_file_path)
return motif_list
def getAlignmentMotifs( sub_node, bedseq, seqalign):
for node_motif in sub_node:
if node_motif.tag.lower() == BedSeqAlignmentStatsCommStruct.MOTIF_TAG:
start = CommStruct.getAttributeAsint( node_motif, BedSeqAlignmentStatsCommStruct.MOTIF_START_ATT)
end = CommStruct.getAttributeAsint( node_motif, BedSeqAlignmentStatsCommStruct.MOTIF_END_ATT)
name = CommStruct.getAttribute( node_motif, BedSeqAlignmentStatsCommStruct.MOTIF_NAME_ATT)
motif_id = CommStruct.getAttribute( node_motif, BedSeqAlignmentStatsCommStruct.MOTIF_ID_ATT, False)
consensus = CommStruct.getAttribute( node_motif, BedSeqAlignmentStatsCommStruct.MOTIF_CONSENSUS_ATT, False)
nb_species = CommStruct.getAttributeAsint( node_motif, BedSeqAlignmentStatsCommStruct.MOTIF_NBSPECIES_ATT, False)
strand = CommStruct.getAttribute( node_motif, BedSeqAlignmentStatsCommStruct.MOTIF_STRAND_ATT, False)
offset = CommStruct.getAttributeAsint( node_motif, BedSeqAlignmentStatsCommStruct.MOTIF_OFFSET_ATT)
score = CommStruct.getAttributeAsfloat( node_motif, BedSeqAlignmentStatsCommStruct.MOTIF_SCORE_ATT)
# Retrieve the PWM of the motif
pwm_s1 = CommStruct.getAttribute( node_motif, BedSeqAlignmentStatsCommStruct.MOTIF_PWM_ATT)
if pwm_s1 != None and len( pwm_s1) > 0:
pwm_matrix = {}
pwm_s2 = pwm_s1.split(";")
for line in pwm_s2:
pwm_s3 = line.split(":")
if len( pwm_s3) > 1:
pwm_s4 = pwm_s3[1].split()
try:
length = 0
for value in pwm_s4:
length += 1
if not pwm_matrix.has_key( pwm_s3[0]):
pwm_matrix[ pwm_s3[0]]=[]
pwm_matrix[ pwm_s3[0]].append( int( value))
except ValueError, val_exce:
raise ParsingException( "BedSeqAlignmentStatsCommStruct.getAlignmentMotifs : Unable to get integer value for Motif '" + BedSeqAlignmentStatsCommStruct.MOTIF_PWM_ATT + "' attributes. From:\n\t---> " + str( val_exce))
pwm = PWM()
pwm.matrix = pwm_matrix
pwm.totalLength = length
pwm.nbSequences = nb_species
else:
pwm = None
if start != None and end != None and name != None:
motif = Motif( start, end, name, pwm)
motif.offset = offset
motif.score = score
if consensus != None:
motif.consensus = consensus
if motif_id != None:
motif.id = motif_id
if strand != None:
motif.strand = strand
seqalign.addMotif( motif)
else:
raise ParsingException( "BedSeqAlignmentStatsCommStruct.getAlignmentMotifs : The motifs of the alignment of '" + bedseq.toString() + "' is missing required attributes")
else:
raise ParsingException( "BedSeqAlignmentStatsCommStruct.getAlignmentMotifs : The motifs of the alignment of '" + bedseq.toString() + "' contains an unauthorized element : '" + node_motif.tag.lower() + "'")
def getMotif(node_motif, comm_struct):
name = CommStruct.getAttribute(
node_motif, MotifStatisticsCommStruct.MOTIF_NAME_ATT)
consensus = CommStruct.getAttribute(
node_motif, MotifStatisticsCommStruct.MOTIF_CONSENSUS_ATT, False)
if name != None:
motif = Motif(0, 0, name, None)
if consensus != None:
motif.consensus = consensus
comm_struct.addMotif(motif)
statistics = MotifStatisticsCommStruct.getMotifStatistics(
node_motif, motif)
comm_struct.addMotifStatistics(motif, statistics)
else:
raise ParsingException(
"MotifStatisticsCommStruct.getAlignmentMotifs : The motif is missing required attribute 'name'"
)
def getMotifMatrices(self, motif_name_list, database_file_path,
database_format):
motif_list = []
for name in motif_name_list:
motif_list.append(Motif(0, 0, name, None))
if database_format == "transfac" or database_format == "tf":
MotifUtils.getMotifsPWMFromJasparTF(motif_list, database_file_path)
elif database_format == "meme":
MotifUtils.getMotifsPWMFromMeme(motif_list, database_file_path)
else:
raise ExecutionException(
"CompareIdentifiedMotifsProcessor.getMotifMatrices : Unknown database format : "
+ database_format)
return motif_list
def extendBlock(self, index_start, index_end, ratio, left_limit, pwm):
# try to extend on the right
# note: since the intervals are semi-open, the value at position index_end was not
# taken into account in the previous ratio. That's why, extending on the right means
# adding the value at position index_end
finished = False
while not finished and index_end <= pwm.totalLength - 1:
if pwm.ratioMatrix[Constants.MAX_INDEX][index_end] >= 0:
new_ratio = self.computeBlockRatio(index_start, index_end + 1,
pwm)
if new_ratio < self.windowConservationLimit:
finished = True
else:
index_end += 1
else:
finished = True
# try to extend on the left but not more then left_limit
finished = False
while not finished and index_start >= 1 and index_start > left_limit:
if pwm.ratioMatrix[Constants.MAX_INDEX][index_start - 1] >= 0:
new_ratio = self.computeBlockRatio(index_start - 1, index_end,
pwm)
if new_ratio < self.windowConservationLimit:
finished = True
else:
index_start -= 1
else:
finished = True
# build the resulting block
block_pwm = pwm.getPWM(index_start, index_end)
block = Motif(index_start, index_end, "", block_pwm)
return block