def histoneFeaturedVector(path, percentage, genome, splitGenome, typeOfFill, percentageOfOverPosition, histoneMarks_by_type): #tipeOfFill sera binario o decimal y percentageOfOverPosition is a percentagen needed to keep in a group
hmFolders = glob(path+"*/")
hmDict = {}
if histoneMarks_by_type == "together":
hmDict["histone_marks"] = get_vector(genome, splitGenome)
if histoneMarks_by_type == "unique":
for hmFolder in hmFolders:
hmDict[hmFolder.split("/")[-2].replace("-human","")] = get_vector(genome, splitGenome)
for hmFolder in hmFolders:
currentHM = hmFolder.split("/")[-2].replace("-human","")
if histoneMarks_by_type == "together":
currentHM = "histone_marks"
#getting the current features to use
features = intersect(hmFolder,percentage)
#getting dictionary of zeros for each portion of the genome
for feature in features:
init = int(feature[1]/(splitGenome*1000))
end = int(feature[2]/(splitGenome*1000))
#percent init ocuppancy = (len(feature_init) * 100)/len(feature)
#len(feature) = feature[2]-feature[1]
#len(feature_init) = ((init_position+1)*(kb*1000)) - feature[1]
percent_init_occupancy = ((((init+1)*(splitGenome*1000.))-feature[1])*100)/(feature[2]-feature[1])
#percent end ocuppancy = (len(feature_end) * 100)/len(feature)
#len(feature) = feature[2]-feature[1]
#len(feature_end) = feature[2] -((end_position)*(kb*1000))
percent_end_occupancy = ((feature[2]-(end*(splitGenome*1000.)))*100)/(feature[2]-feature[1])
i = init
while i <= end:
to_mark = True # to know if I need to mark that position
if i == init:
if percent_init_occupancy < percentageOfOverPosition:
to_mark == False
if i == end:
if percent_end_occupancy < percentageOfOverPosition:
to_mark == False
if to_mark == True:
aux = hmDict[currentHM][feature[0]]
if typeOfFill == "decimal":
aux[i] +=1
if typeOfFill == "binary":
aux[i] = 1
hmDict[currentHM][feature[0]] = aux
i+=1
return hmDict
def geneVector(genome, splitGenome, GTF, typeOfFill, percentageOfOverPosition):
#getting gene vector for each chr (its a dict)
vector = get_vector(genome, splitGenome)
gtfGene = []
gtf = open(GTF,"r")
for line in gtf:
splittedLine = line.split("\t")
geneID = splittedLine[8].split("\"")[1]
if "spikein exon" not in line:
if (splittedLine[2] == "gene" or "ENSG" not in geneID):
#getting gene coords
c1 = float(splittedLine[3])
c2 = float(splittedLine[4])
init = int(c1/(splitGenome*1000))
end = int(c2/(splitGenome*1000))
#percent init ocuppancy = (len(feature_init) * 100)/len(feature)
#len(feature) = feature[2]-feature[1]
#len(feature_init) = ((init_position+1)*(kb*1000)) - feature[1]
percent_init_occupancy = ((((init+1)*(splitGenome*1000.))-c1)*100)/(c2-c1)
#percent end ocuppancy = (len(feature_end) * 100)/len(feature)
#len(feature) = feature[2]-feature[1]
#len(feature_end) = feature[2] -((end_position)*(kb*1000))
percent_end_occupancy = ((c2-(end*(splitGenome*1000.)))*100)/(c2-c1)
i = init
while i <= end:
to_mark = True # to know if I need to mark that position
if i == init:
if percent_init_occupancy < percentageOfOverPosition:
to_mark == False
if i == end:
if percent_end_occupancy < percentageOfOverPosition:
to_mark == False
if to_mark == True:
aux = vector[splittedLine[0]]
if typeOfFill == "decimal":
aux[i] +=1
if typeOfFill == "binary":
aux[i] = 1
vector[splittedLine[0]] = aux
i+=1
return vector
def getFeaturedVector(path, percentage, genome, splitGenome, typeOfFill, percentageOfOverPosition): #tipeOfFill sera binario o decimal y percentageOfOverPosition is a percentagen needed to keep in a group #getting the current features to use features = intersect(path,percentage) #getting dictionary of zeros for each portion of the genome dictZeroVectors = get_vector(genome, splitGenome) for feature in features: init = int(feature[1]/(splitGenome*1000)) end = int(feature[2]/(splitGenome*1000)) #percent init ocuppancy = (len(feature_init) * 100)/len(feature) #len(feature) = feature[2]-feature[1] #len(feature_init) = ((init_position+1)*(kb*1000)) - feature[1] percent_init_occupancy = ((((init+1)*(splitGenome*1000.))-feature[1])*100)/(feature[2]-feature[1]) #percent end ocuppancy = (len(feature_end) * 100)/len(feature) #len(feature) = feature[2]-feature[1] #len(feature_end) = feature[2] -((end_position)*(kb*1000)) percent_end_occupancy = ((feature[2]-(end*(splitGenome*1000.)))*100)/(feature[2]-feature[1]) i = init while i <= end: to_mark = True # to know if I need to mark that position if i == init: if percent_init_occupancy < percentageOfOverPosition: to_mark == False if i == end: if percent_end_occupancy < percentageOfOverPosition: to_mark == False if to_mark == True: aux = dictZeroVectors[feature[0]] if typeOfFill == "decimal": aux[i] +=1 if typeOfFill == "binary": aux[i] = 1 dictZeroVectors[feature[0]] = aux i+=1 return dictZeroVectors
def FIMOVector(fimopath, genome, splitGenome, typeOfFill,
percentageOfOverPosition, fimo_filter):
fimoDict = {} #TFBM = NSITES
fimoMotifs = open("./dat/HOCOMOCOv11_full_HUMAN_mono_meme_format.meme",
"r")
currentMotif = ""
for line in fimoMotifs:
if "MOTIF" in line:
currentMotif = line[:-1].replace("MOTIF ", "")
if "nsites=" in line:
fimoDict[currentMotif] = int(line[:-1].split("nsites= ")[-1])
fimoMotifs.close
folders = glob(fimopath + "/*/")
dictZeroVectors = get_vector(genome, splitGenome)
for folder in folders:
file = open(folder + "fimo.txt", "r")
for line in file:
if line[0] != "#":
splt = line[:-1].split("\t")
if fimoDict[splt[0]] >= fimo_filter:
chr = splt[2]
#print splt[1],splt[3], splt[4]
init = int(int(splt[3]) / (splitGenome * 1000))
end = int(int(splt[4]) / (splitGenome * 1000))
#percent init ocuppancy = (len(feature_init) * 100)/len(feature)
#len(feature) = feature[2]-feature[1]
#len(feature_init) = ((init_position+1)*(kb*1000)) - feature[1]
percent_init_occupancy = (
(((init + 1) * (splitGenome * 1000.)) - int(splt[3])) *
100) / (int(splt[4]) - int(splt[3]))
#percent end ocuppancy = (len(feature_end) * 100)/len(feature)
#len(feature) = feature[2]-feature[1]
#len(feature_end) = feature[2] -((end_position)*(kb*1000))
percent_end_occupancy = (
(int(splt[4]) - (end * (splitGenome * 1000.))) *
100) / (int(splt[4]) - int(splt[3]))
#if init!=end:
# print init, end, percent_init_occupancy, percent_end_occupancy
i = init
while i <= end:
to_mark = True # to know if I need to mark that position
if i == init:
if percent_init_occupancy < percentageOfOverPosition:
to_mark == False
if i == end:
if percent_end_occupancy < percentageOfOverPosition:
to_mark == False
if to_mark == True:
aux = dictZeroVectors[chr]
if typeOfFill == "decimal":
aux[i] += 1
if typeOfFill == "binary":
aux[1] = 1
dictZeroVectors[chr] = aux
i += 1
file.close()
return dictZeroVectors
def tfFeaturedVector(
path, percentage, genome, splitGenome, typeOfFill, percentageOfOverPosition
): #tipeOfFill sera binario o decimal y percentageOfOverPosition is a percentagen needed to keep in a group
#reading the list of TFs
tfFile = open("./dat/TFs.csv", "r")
tfList = []
for line in tfFile:
if line[0] != "#": #if the current line is not a comment
aux = line.split(",")
if aux[3] == "Yes":
tfList.append(aux[1])
tfFile.close()
tfFolders = glob(path + "*/")
dictZeroVectors = get_vector(genome, splitGenome)
for tfFolder in tfFolders:
currentTF = tfFolder.split("/")[-2].replace("-human", "")
if currentTF in tfList:
#getting the current features to use
features = intersect(tfFolder, percentage)
#getting dictionary of zeros for each portion of the genome
for feature in features:
init = int(feature[1] / (splitGenome * 1000))
end = int(feature[2] / (splitGenome * 1000))
#percent init ocuppancy = (len(feature_init) * 100)/len(feature)
#len(feature) = feature[2]-feature[1]
#len(feature_init) = ((init_position+1)*(kb*1000)) - feature[1]
percent_init_occupancy = (((
(init + 1) * (splitGenome * 1000.)) - feature[1]) *
100) / (feature[2] - feature[1])
#percent end ocuppancy = (len(feature_end) * 100)/len(feature)
#len(feature) = feature[2]-feature[1]
#len(feature_end) = feature[2] -((end_position)*(kb*1000))
percent_end_occupancy = ((feature[2] -
(end * (splitGenome * 1000.))) *
100) / (feature[2] - feature[1])
i = init
while i <= end:
to_mark = True # to know if I need to mark that position
if i == init:
if percent_init_occupancy < percentageOfOverPosition:
to_mark == False
if i == end:
if percent_end_occupancy < percentageOfOverPosition:
to_mark == False
if to_mark == True:
aux = dictZeroVectors[feature[0]]
if typeOfFill == "decimal":
aux[i] += 1
if typeOfFill == "binary":
aux[1] = 1
dictZeroVectors[feature[0]] = aux
i += 1
return dictZeroVectors
def transcriptVector(genome, gtf, RNAseqFolder, splitGenome, typeOfFill,
percentageOfOverPosition):
#vector of genome fragments
vector = get_vector(
genome,
splitGenome) #in the reallity this is a dict of chr: [vector of zeros]
vectorAux = get_vector(
genome, splitGenome
) #an aux to know sum(fpkm)/how many sums where done in this position
#transcript Dictionary
transDict = transcriptDict(gtf) #transcript: [init,end,chr, gene]
#looking for tsv files where fpkm is annoted
tsvs = glob(RNAseqFolder + "/*.tsv")
for tsv in tsvs:
tsvFile = open(tsv, "r")
aux = 0 # to avoid the first line
aux2 = None
for line in tsvFile:
if aux != 0:
split = line.split("\t")
transcriptID = split[0]
fpkm = float(split[6])
if transcriptID in transDict:
coords = transDict[transcriptID]
#now we need to sum the fpkm to the current position and also we will add 1 to the positions in vectoraux
#getting gene coords
c1 = float(coords[0])
c2 = float(coords[1])
init = int(c1 / (splitGenome * 1000))
end = int(c2 / (splitGenome * 1000))
#percent init ocuppancy = (len(feature_init) * 100)/len(feature)
#len(feature) = feature[2]-feature[1]
#len(feature_init) = ((init_position+1)*(kb*1000)) - feature[1]
percent_init_occupancy = ((
((init + 1) *
(splitGenome * 1000.)) - c1) * 100) / (c2 - c1)
#percent end ocuppancy = (len(feature_end) * 100)/len(feature)
#len(feature) = feature[2]-feature[1]
#len(feature_end) = feature[2] -((end_position)*(kb*1000))
percent_end_occupancy = (
(c2 - (end * (splitGenome * 1000.))) * 100) / (c2 - c1)
i = init
while i <= end:
to_mark = True # to know if I need to mark that position
if i == init:
if percent_init_occupancy < percentageOfOverPosition:
to_mark == False
if i == end:
if percent_end_occupancy < percentageOfOverPosition:
to_mark == False
if to_mark == True:
aux2 = vector[transDict[transcriptID][
2]] # <------------------- Este splitted es el cromosoma actual
if typeOfFill == "decimal":
aux2[i] += 1
if typeOfFill == "binary":
aux2[i] = 1
if typeOfFill == "average":
#we will use an auxiliar vector to know how man times it was modified
vectorAux[transDict[transcriptID][2]][i] += 1
aux2[i] += fpkm
#print fpkm
vector[transDict[transcriptID][2]] = aux2
i += 1
aux = 1
tsvFile.close()
#Now we exit from the tsv loop, so is time to divide if typeOfFill was average
if typeOfFill == "average":
for vect in vector.items():
i = 0
for fpkm in vect[1]:
if fpkm != 0:
vector[vect[0]][i] /= vectorAux[vect[0]][i]
i += 1
return vector