def KNNprotein(fastas, **kw):
trainFile = kw['train']
labelFile = kw['label']
if os.path.exists(labelFile) == False:
print('Error: the label file does not exist.')
sys.exit(1)
if trainFile == None or labelFile == None:
print(
'Error: please specify the directory of train file ["--train"] and the label file ["--label"]'
)
sys.exit(1)
trainData = readFasta.readFasta(trainFile)
with open(labelFile) as f:
records = f.readlines()
myLabel = {}
for i in records:
array = i.rstrip().split() if i.strip() != '' else None
myLabel[array[0]] = int(array[1])
myLabelSets = list(set(myLabel.values()))
if len(trainData) != len(myLabel):
print('ERROR: the inconsistent sample number in train and label file.')
sys.exit(1)
kValues = [
0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.10, 0.11, 0.12,
0.13, 0.14, 0.15, 0.16, 0.17, 0.18, 0.19, 0.20, 0.21, 0.22, 0.23, 0.24,
0.25, 0.26, 0.27, 0.28, 0.29, 0.30
]
kNum = []
for i in kValues:
kNum.append(math.ceil(len(trainData) * i))
encodings = []
header = ['#']
for k in kValues:
for l in myLabelSets:
header.append('Top' + str(k) + '.label' + str(l))
encodings.append(header)
for i in fastas:
name, sequence = i[0], re.sub('[^ARNDCQEGHILKMFPSTWYV-]', '', i[1])
code = [name]
mySimilarity = []
for j in range(len(trainData)):
if name != trainData[j][0]:
mySimilarity.append([
myLabel[trainData[j][0]],
CalculateSimilarity(
re.sub('[^ARNDCQEGHILKMFPSTWYV]', '', trainData[j][1]),
sequence)
])
mySimilarity = np.array(mySimilarity)
mySimilarity = mySimilarity[np.lexsort(-mySimilarity.T)]
for j in kNum:
code = code + CalculateContent(mySimilarity, j, myLabelSets)
encodings.append(code)
return encodings
def get_feature(protein_fasta_file, feature_type, output_dir):
fastas = readFasta.readFasta(protein_fasta_file)
userDefinedOrder = 'ACDEFGHIKLMNPQRSTVWY'
myAAorder = {
'alphabetically': 'ACDEFGHIKLMNPQRSTVWY',
'polarity': 'DENKRQHSGTAPYVMCWIFL',
'sideChainVolume': 'GASDPCTNEVHQILMKRFYW',
'userDefined': userDefinedOrder
}
myOrder = 'ACDEFGHIKLMNPQRSTVWY'
kw = {'order': myOrder}
myFun = f"{feature_type}.{feature_type}(fastas, **kw)"
print('Descriptor type: ' + feature_type)
encodings = eval(myFun)
outFile = f'{output_dir}{feature_type}.tsv'
saveCode.savetsv(encodings, outFile)
return
if os.path.exists(outDir) == False:
os.mkdir(outDir)
for i in fastas:
name, sequence = re.sub('\|', '', i[0]), i[1]
with open(name + '.txt', 'w') as f:
f.write('>' + name + '\n' + sequence + '\n')
myCmd = psipred + ' ' + name + '.txt'
if os.path.exists(outDir + '/' + name + '.ss2') == False:
os.system(myCmd)
os.remove(name + '.txt')
os.remove(name + '.ss')
os.remove(name + '.horiz')
shutil.move(name + '.ss2', outDir)
return outDir
if __name__ == '__main__':
parser = argparse.ArgumentParser(
usage="it's usage tip.",
description="generate protein secondary structure profile")
parser.add_argument("--file",
required=True,
help="protein sequence file in fasta format")
parser.add_argument("--psipred", help="the path of psipred program")
args = parser.parse_args()
psipred = args.psipred if args.psipred != None else 'runpsipred'
fastas = readFasta.readFasta(args.file)
outputDir = generateSecondaryStructure(fastas, 'out', psipred)
print('The predicted secodnary structure are stored in: ' + outputDir)
myDistance.append([
myLabel[trainData[j][0]],
CalculateDistance(trainData[j][1], sequence)
])
myDistance = np.array(myDistance)
myDistance = myDistance[np.lexsort(myDistance.T)]
for j in kNum:
code = code + CalculateContent(myDistance, j, myLabelSets)
encodings.append(code)
return encodings
fastas = readFasta.readFasta("data_test.txt")
kw = {
'path': "D:\\xw\\3_特征提取\\KNN\\codes\\codes",
'train': "data_test.txt",
'label': "label_test.txt",
'order': 'ACDEFGHIKLMNPQRSTVWY'
}
#kw= {'path': "D:\\xw\\S-sulfenylation_特征提取\\KNN\\KNN0710\\codes",'train':"data_train.txt",'label':"label_train.txt",'order':'ACDEFGHIKLMNPQRSTVWY'}
data_KNN = KNNpeptide(fastas, **kw)
data_raw = data_KNN[1:]
data_new = np.matrix(data_raw)
data_knn = data_new[:, 1:]
column = data_knn.shape[1]
data_knn_feature = data_knn[:, np.array(range(1, column, 2))]
test = pd.DataFrame(data=data_knn_feature)
encodings.append(header)
for i in fastas:
name, sequence = i[0], i[1]
code = [name]
for aa in sequence:
if aa == '-':
for j in AAindex:
code.append(0)
continue
for j in AAindex:
code.append(j[index[aa]])
encodings.append(code)
return encodings
fastas = readFasta.readFasta(r"F:\python\KNN\test_A.txt")
kw = {
'path': r"F:\python\KNN",
'train': r"F:\python\KNN\test_A.txt",
'label': r"F:\python\KNN\label_A.txt",
'order': 'ACDEFGHIKLMNPQRSTVWY'
}
result = AAINDEX(fastas, **kw)
data = result[1:]
data_new = np.matrix(data)
data_AAindex = data_new[:, 1:]
data_AAindex_end = pd.DataFrame(data=data_AAindex)
data_AAindex_end.to_csv('AAindex.csv')
def KNNpeptide(fastas, **kw):
trainFile = kw['train']
labelFile = kw['label']
if trainFile == None or labelFile == None:
print('Error: please specify the directory of train file ["--train"] and the label file ["--label"]')
sys.exit(1)
if os.path.exists(labelFile) == False:
print('Error: the label file does not exist.')
sys.exit(1)
trainData = readFasta.readFasta(trainFile)
with open(labelFile) as f:
records = f.readlines()
myLabel = {}
countme=1
for i in records:
#print(str(countme))
array = i.rstrip().split('\t') if i.strip('\t') != '' else None
if array[0] in myLabel:
print (array[0])
myLabel[array[0]] = int(array[1])
#print(array[0])
#print(array[1])
countme=countme+1
myLabelSets = list(set(myLabel.values()))
if len(trainData) != len(myLabel):
print(len(myLabel))
print(len(trainData))
print(len(myLabel))
print('ERROR: inconsistent sample number between train and label file.')
sys.exit(1)
#這邊是照比率的話
kValues = [0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.10]
kNum = []
for i in kValues:
kNum.append(math.ceil(len(trainData) * i))
#kNum = [2,4,8,16,32,64]
encodings = []
header = ['#']
for k in kNum:
for l in myLabelSets:
header.append('Top' + str(k) + '.label' + str(l))
encodings.append(header)
for i in fastas:
name, sequence = i[0], i[1]
code = [name]
myDistance = []
#print(name)
#print(sequence)
for j in range(len(trainData)):
if name != trainData[j][0]:
myDistance.append([myLabel[trainData[j][0]], CalculateDistance(trainData[j][1], sequence)])
#print("hello")
myDistance = np.array(myDistance)
myDistance = myDistance[np.lexsort(myDistance.T)]
for j in kNum:
code = code + CalculateContent(myDistance, j, myLabelSets)
encodings.append(code)
return encodings
return distance
def CalculateContent(myDistance, j, myLabelSets):
content = []
myDict = {}
for i in myLabelSets:
myDict[i] = 0
for i in range(j):
myDict[myDistance[i][0]] = myDict[myDistance[i][0]] + 1
for i in myLabelSets:
content.append(myDict[myLabelSets[i]] / j)
return content
ffastas = readFasta.readFasta(r"F:\python\KNN\test_A.txt")
kw = {
'path': r"F:\python\KNN",
'train': r"F:\python\KNN\test_A.txt",
'label': r"F:\python\KNN\label_A.txt",
'order': 'ACDEFGHIKLMNPQRSTVWY'
}
trainFile = kw['train']
labelFile = kw['label']
if trainFile == None or labelFile == None:
print(
'Error: please specify the directory of train file ["--train"] and the label file ["--label"]'
)
sys.exit(1)
if os.path.exists(labelFile) == False: