def EIGAs(dataBaseDir, parsedDataFile, parseData, saveParsedData, \
kappa, gapOpen, gapExtend, numProcessors, classificationFile):
# Open Log File
if not os.path.exists('../logs/'):
os.makedirs('../logs/')
logFile = open('../logs/runlog','w')
print('EIGAs -> Run details in '+'../logs/runlog')
# Parse Proteins
print('EIGAs -> Collecting and Encoding the Proteins')
ticParse = time.time()
if parseData:
logFile.write('Processing files from '+dataBaseDir+'\n')
tic = time.time()
Prot = EIGAs_utils.readProteins(logFile, dataBaseDir, classificationFile)
logFile.write('\t required '+str(time.time()-tic)+' seconds\n')
# Encode each chain as a list of eigenvalues
tic = time.time()
Prot = EIGAs_utils.encode(Prot, kappa, logFile)
logFile.write('\t required '+str(time.time()-tic)+' seconds\n')
# Save the datafile
if saveParsedData:
logFile.write('Saving parsed data in '+parsedDataFile+'\n')
tic = time.time()
pklFile = open(parsedDataFile, 'wb')
cPickle.dump(Prot, pklFile)
pklFile.close()
logFile.write('\t required '+str(time.time()-tic)+' seconds\n')
else:
# Load the datafile
if os.path.exists(parsedDataFile):
logFile.write('Loading parsed data from '+parsedDataFile+'\n')
tic = time.time()
pklFile = open(parsedDataFile, 'rb')
Prot = cPickle.load(pklFile)
pklFile.close()
logFile.write('\t required '+str(time.time()-tic)+' seconds\n')
else:
logFile.write('No data file found\n')
print('\t\t required '+str(time.time()-ticParse)+' seconds')
# Perform pairwise alignments
print('EIGAs -> Conducting pairwise alignments')
logFile.write('\n ***** EIGAs Beginning Pairwise Alignments *****\n')
ticAlign = time.time()
if numProcessors > 1:
pool = Pool(processes=numProcessors)
alignment = {}
for (prtName1, prtName2) in itertools.combinations(Prot, 2):
# Check to make sure we have legitimate chains
if Prot[prtName1].model[0].chain[0].legitChain and \
Prot[prtName2].model[0].chain[0].legitChain:
# Send off to align with DP
if numProcessors > 1:
alignment[(prtName1, prtName2, 'score')], \
alignment[(prtName1, prtName2, 'seq1')], \
alignment[(prtName1, prtName2, 'lenSeq1')], \
alignment[(prtName1, prtName2, 'seq2')], \
alignment[(prtName1, prtName2, 'lenSeq2')] = \
pool.apply_async(EIGAs_utils.align, \
(Prot[prtName1].model[0].chain[0].lambdaVal, \
Prot[prtName2].model[0].chain[0].lambdaVal, \
gapOpen, gapExtend) ).get()
elif numProcessors == 1:
alignment[(prtName1, prtName2, 'score')], \
alignment[(prtName1, prtName2, 'seq1')], \
alignment[(prtName1, prtName2, 'lenSeq1')], \
alignment[(prtName1, prtName2, 'seq2')], \
alignment[(prtName1, prtName2, 'lenSeq2')] = \
EIGAs_utils.align(\
Prot[prtName1].model[0].chain[0].lambdaVal, \
Prot[prtName2].model[0].chain[0].lambdaVal, \
gapOpen, gapExtend)
else:
logFile.write('Error, numProcessors is unknown\n')
return {}
# normalize the score
alignment[(prtName1, prtName2, 'score')] = \
alignment[(prtName1, prtName2, 'score')] / (0.5 * \
len(Prot[prtName1].model[0].chain[0].lambdaVal) + \
len(Prot[prtName2].model[0].chain[0].lambdaVal) )
# Calculate the sCMO score
alignment[(prtName1, prtName2, 'sCMO')] = \
EIGAs_utils.calc_sCMO(\
Prot[prtName1].model[0].chain[0].contactMat, \
alignment[(prtName1, prtName2, 'seq1')],\
Prot[prtName2].model[0].chain[0].contactMat,\
alignment[(prtName1, prtName2, 'seq2')])
# store the symmetric comparison
alignment[(prtName2, prtName1, 'score')] = \
alignment[(prtName1, prtName2, 'score')]
alignment[(prtName2, prtName1, 'seq1')] = \
alignment[(prtName1, prtName2, 'seq2')]
alignment[(prtName2, prtName1, 'lenSeq1')] = \
alignment[(prtName1, prtName2, 'lenSeq2')]
alignment[(prtName2, prtName1, 'seq2')] = \
alignment[(prtName1, prtName2, 'seq1')]
alignment[(prtName2, prtName1, 'lenSeq2')] = \
alignment[(prtName1, prtName2, 'lenSeq1')]
alignment[(prtName2, prtName1, 'sCMO')] = \
alignment[(prtName1, prtName2, 'sCMO')]
# log the pairwise comparison
logFile.write(str(prtName1)+', '+str(prtName2)+' has value '+\
str(alignment[(prtName1, prtName2, 'score')])+'\n')
# Close the multiprocessing pool
if numProcessors > 1:
pool.close()
pool.join()
logFile.write('\t required '+str(time.time()-ticAlign)+' seconds over '\
+str(numProcessors)+' processors\n')
logFile.write('\n ***** EIGAs Done With Pairwise Alignments *****\n')
print('\t\t required '+str(time.time()-ticAlign)+' seconds')
# Assess the alignments
print('EIGAs -> Assessing the Alignments')
logFile.write('\nAssessing the alignments\n')
ticAssess = time.time()
assessment={}
assessment['numCorrect'], assessment['TPR'], assessment['FPR'], \
assessment['ROCarea'], assessment['partitionSize'], \
assessment['maxScore'], assessment['minScore'], \
assessment['bestDiscernmentVal'], assessment['bestTPR'], \
assessment['bestFPR'] = EIGAs_utils.assessAlignment(Prot, alignment)
assessment['numProteins'] = len(Prot)
logFile.write('\t required '+str(time.time()-tic)+' seconds\n')
logFile.write('\t '+str(assessment['numCorrect'])+'/'+str(len(Prot))+\
' were correct as measured by best score\n')
logFile.write('\t ROC curve area was '+str(assessment['ROCarea'])+'\n')
logFile.write('\t minScore <= bestDiscernment <= maxScore were \n')
logFile.write('\t\t'+str(assessment['minScore'])+' <= '+\
str(assessment['bestDiscernmentVal'])+' <= '+\
str(assessment['maxScore'])+'\n')
logFile.write('\t best TPR and FPR were '+str(assessment['bestTPR'])+\
' and '+str(assessment['bestFPR'])+'\n')
dist = sqrt((1-assessment['bestTPR'])**2+assessment['bestFPR']**2)
logFile.write('\t\t distance to perfect (0,1) on ROC curve was '+\
str(dist)+'\n')
print('\t\t required '+str(time.time()-ticAssess)+' seconds')
# Close Log File
logFile.close()
return assessment
return Math.acos((a[1]*b[1] + a[2]*b[2] + a[0]*b[0])/(Math.pow(a[1]*a[1]+a[0]*a[0]+a[2]*a[2],0.5)*Math.pow(b[1]*b[1]+b[0]*b[0] + b[2]*b[2],0.5)))
def normalize(a):
max = None
for item in a:
size = Math.pow(item[0]*item[0] + item[1]*item[1] + item[2]*item[2],0.5)
if(max == None or size > max):
max = size
ret = []
for item in a:
ret += [((item[0]+0.0)/max, (item[1]+0.0)/max,(item[2]+0.0)/max)]
return ret
def itemDiv(sc):
ret = []
for i in sc:
r2 = []
for j in i:
r2 += [j / max(sc)[0]]
ret += [r2]
return rets
logFile = open('../logs/runlog','w')
Prot = EIGAs_utils.readProteins(logFile, '../data/Skolnick40/', '../data/SCOP/dir.cla.scop.txt_1.75')
amk = Prot["1AMK"].model[0].chain[0].coords
aw2 = Prot["1AW2"].model[0].chain[0].coords
f = open("results.txt",'w')
#print score(characteristics(amk,15),characteristics(aw2,15))
for i in score(characteristics(amk,15),characteristics(aw2,15)):
f.write(str(i))
f.close()
