def PAAC(fastas, lambdaValue=2, w=0.05, **kw):
if checkFasta.minSequenceLengthWithNormalAA(fastas) < lambdaValue + 1:
print(
'Error: all the sequence length should be larger than the lambdaValue+1: '
+ str(lambdaValue + 1) + '\n\n')
return 0
dataFile = re.sub(
'codes$', '',
os.path.split(os.path.realpath(__file__))
[0]) + r'\data\PAAC.txt' if platform.system() == 'Windows' else re.sub(
'codes$', '',
os.path.split(os.path.realpath(__file__))[0]) + '/data/PAAC.txt'
with open(dataFile) as f:
records = f.readlines()
AA = ''.join(records[0].rstrip().split()[1:])
AADict = {}
for i in range(len(AA)):
AADict[AA[i]] = i
AAProperty = []
AAPropertyNames = []
for i in range(1, len(records)): #len(records)is4
array = records[i].rstrip().split(
) if records[i].rstrip() != '' else None
AAProperty.append([float(j) for j in array[1:]])
AAPropertyNames.append(array[0])
AAProperty1 = []
for i in AAProperty:
meanI = sum(i) / 20
fenmu = math.sqrt(sum([(j - meanI)**2 for j in i]) / 20)
AAProperty1.append([(j - meanI) / fenmu for j in i])
encodings = []
header = ['#']
for aa in AA:
header.append('Xc1.' + aa)
for n in range(1, lambdaValue + 1):
header.append('Xc2.lambda' + str(n))
encodings.append(header)
for i in fastas:
name, sequence = i[0], re.sub('-', '', i[1])
code = [name]
theta = []
for n in range(1, lambdaValue + 1):
theta.append(
sum([
Rvalue(sequence[j], sequence[j + n], AADict, AAProperty1)
for j in range(len(sequence) - n)
]) / (len(sequence) - n))
myDict = {}
for aa in AA:
myDict[aa] = sequence.count(aa)
code = code + [myDict[aa] / (1 + w * sum(theta)) for aa in AA]
code = code + [(w * j) / (1 + w * sum(theta)) for j in theta]
encodings.append(code)
return encodings
def SOCNumber(fastas, nlag=30, **kw):
if checkFasta.minSequenceLengthWithNormalAA(fastas) < nlag + 1:
print('Error: all the sequence length should be larger than the nlag+1: ' + str(nlag + 1) + '\n\n')
return 0
dataFile = re.sub('codes$', '', os.path.split(os.path.realpath(__file__))[0]) + r'\data\Schneider-Wrede.txt' if platform.system() == 'Windows' else re.sub('codes$', '', os.path.split(os.path.realpath(__file__))[0]) + '/data/Schneider-Wrede.txt'
dataFile1 = re.sub('codes$', '', os.path.split(os.path.realpath(__file__))[0]) + r'\data\Grantham.txt' if platform.system() == 'Windows' else re.sub('codes$', '', os.path.split(os.path.realpath(__file__))[0]) + '/data/Grantham.txt'
AA = 'ACDEFGHIKLMNPQRSTVWY'
AA1 = 'ARNDCQEGHILKMFPSTWYV'
DictAA = {}
for i in range(len(AA)):
DictAA[AA[i]] = i
DictAA1 = {}
for i in range(len(AA1)):
DictAA1[AA1[i]] = i
with open(dataFile) as f:
records = f.readlines()[1:]
AADistance = []
for i in records:
array = i.rstrip().split()[1:] if i.rstrip() != '' else None
AADistance.append(array)
AADistance = np.array(
[float(AADistance[i][j]) for i in range(len(AADistance)) for j in range(len(AADistance[i]))]).reshape((20, 20))
with open(dataFile1) as f:
records = f.readlines()[1:]
AADistance1 = []
for i in records:
array = i.rstrip().split()[1:] if i.rstrip() != '' else None
AADistance1.append(array)
AADistance1 = np.array(
[float(AADistance1[i][j]) for i in range(len(AADistance1)) for j in range(len(AADistance1[i]))]).reshape(
(20, 20))
encodings = []
header = ['#']
for n in range(1, nlag + 1):
header.append('Schneider.lag' + str(n))
for n in range(1, nlag + 1):
header.append('gGrantham.lag' + str(n))
encodings.append(header)
for i in fastas:
name, sequence = i[0], re.sub('-', '', i[1])
code = [name]
for n in range(1, nlag + 1):
code.append(sum(
[AADistance[DictAA[sequence[j]]][DictAA[sequence[j + n]]] ** 2 for j in range(len(sequence) - n)]) / (
len(sequence) - n))
for n in range(1, nlag + 1):
code.append(sum([AADistance1[DictAA1[sequence[j]]][DictAA1[sequence[j + n]]] ** 2 for j in
range(len(sequence) - n)]) / (len(sequence) - n))
encodings.append(code)
return encodings
def NMBroto(fastas,
props=[
'CIDH920105', 'BHAR880101', 'CHAM820101', 'CHAM820102',
'CHOC760101', 'BIGC670101', 'CHAM810101', 'DAYM780201'
],
nlag=5,
**kw):
if checkFasta.minSequenceLengthWithNormalAA(fastas) < nlag + 1:
print(
'Error: all the sequence length should be larger than the nlag+1: '
+ str(nlag + 1) + '\n\n')
return 0
AA = 'ARNDCQEGHILKMFPSTWYV'
fileAAidx = re.sub('codes$', '',
os.path.split(os.path.realpath(__file__))
[0]) + r'\data\AAidx.txt' if platform.system(
) == 'Windows' else sys.path[0] + '/data/AAidx.txt'
with open(fileAAidx) as f:
records = f.readlines()[1:]
myDict = {}
for i in records:
array = i.rstrip().split('\t')
myDict[array[0]] = array[1:]
AAidx = []
AAidxName = []
for i in props:
if i in myDict:
AAidx.append(myDict[i])
AAidxName.append(i)
else:
print('"' + i + '" properties not exist.')
return None
AAidx1 = np.array([float(j) for i in AAidx for j in i])
AAidx = AAidx1.reshape((len(AAidx), 20))
pstd = np.std(AAidx, axis=1)
pmean = np.average(AAidx, axis=1)
for i in range(len(AAidx)):
for j in range(len(AAidx[i])):
AAidx[i][j] = (AAidx[i][j] - pmean[i]) / pstd[i]
index = {}
for i in range(len(AA)):
index[AA[i]] = i
encodings = []
header = ['#']
for p in props:
for n in range(1, nlag + 1):
header.append(p + '.lag' + str(n))
encodings.append(header)
for i in fastas:
name, sequence = i[0], re.sub('-', '', i[1])
code = [name]
N = len(sequence)
for prop in range(len(props)):
for n in range(1, nlag + 1):
if len(sequence) > nlag:
# if key is '-', then the value is 0
rn = sum([
AAidx[prop][index.get(sequence[j], 0)] *
AAidx[prop][index.get(sequence[j + n], 0)]
for j in range(len(sequence) - n)
]) / (N - n)
else:
rn = 'NA'
code.append(rn)
encodings.append(code)
return encodings
