def main():
if (len(sys.argv) != 3): # if no input
print "ERORR"
print "syntax: python generate_chemaxon_fingerprints.py smiles1 outputprefix"
return
smilesfile1 = sys.argv[1]
outfileprefix = sys.argv[2]
outfileF = outfileprefix + '.fp'
pid = str(os.getpid(
)) # get the process idenifier so that we do not right over the same file.
# read in names from smiles file.
names = []
file1 = open(smilesfile1, 'r')
for line in file1:
print line
name = line.split()[1]
#print line, name
names.append(name)
file1.close()
# if the fp file already exists, just read in the fingerprints.
# otherwise caluclate the fingerprints.
if (os.path.isfile(outfileF)):
print outfileF + " exists."
exit()
else:
fpvec1 = tccalc.get_fp(smilesfile1, outfileF, pid)
def main():
if not (len(sys.argv) == 4 or len(sys.argv) == 5): # if no input
print "ERORR"
print "syntexs: python tanimoto_cal_axon.py -one smiles1 outputprefix"
print " this produces a squere symestric matrix of set1 with itself. "
print "syntexs: python tanimoto_cal_axon.py -two smiles1 smiles2 outputprefix"
print " this produces a rectangular non-symestric matrix of set1 to set2"
return
pid = str(os.getpid()) # get the process idenifier so that we do not right over the same file.
print pid
oneortwo = sys.argv[1]
smilesfile1 = sys.argv[2]
if oneortwo == "-one":
outfileprefix = sys.argv[3]
elif oneortwo == "-two":
smilesfile2 = sys.argv[3]
outfileprefix = sys.argv[4]
else:
print "the frist parameter must be -one or -two."
exit()
outfile1 = outfileprefix +'.1.fp'
fpvec1 = tccalc.get_fp(smilesfile1,outfile1,pid)
if oneortwo == "-one":
fpvec2 = fpvec1
if oneortwo == "-two":
outfile2 = outfileprefix +'.2.fp'
fpvec2 = tccalc.get_fp(smilesfile2,outfile2,pid)
outfileM = outfileprefix +'.tanimoto.matrix'
#print len(fpvec2)
#print len(fpvec1)
#exit(0)
file1 = open(outfileM,'w')
for fp1 in fpvec1:
flag_frist = True
for fp2 in fpvec2:
#print fp1
#print fp2
TC = tccalc.tanimoto(fp1,fp2)
if (flag_frist):
flag_frist = False
else:
file1.write(',')
file1.write('%f' % TC )
file1.write('\n' )
file1.close()
alpha = 0.2
beta = 0.2
outfileM = outfileprefix +'.tversky.'+str(alpha)+'.'+str(beta)+'.matrix'
file1 = open(outfileM,'w')
for fp1 in fpvec1:
flag_frist = True
for fp2 in fpvec2:
TV = tccalc.tversky_index(fp1,fp2,alpha,beta)
if (flag_frist):
flag_frist = False
else:
file1.write(',')
file1.write('%f' % TV )
file1.write('\n' )
file1.close()
def main():
if (len(sys.argv) != 4): # if no input
print "ERORR"
print "syntexs: python best_frist_tc_clusters.py smiles1 outputprefix threshold"
return
smilesfile1 = sys.argv[1]
outfileprefix = sys.argv[2]
tc_threshold = float(sys.argv[3])
outfileF = outfileprefix +'.fp'
pid = str(os.getpid()) # get the process idenifier so that we do not right over the same file.
# read in names from smiles file.
names = []
file1 = open(smilesfile1,'r')
for line in file1:
name = line.split()[1]
#print line, name
names.append(name)
file1.close()
# if the fp file already exists, just read in the fingerprints.
# otherwise caluclate the fingerprints.
if (os.path.isfile(outfileF) ):
print outfileF + "exists."
fpvec1 = []
file1 = open(outfileF)
for line in file1:
fp = line.split()[2]
#print fp
fpvec1.append(fp)
else:
fpvec1 = tccalc.get_fp(smilesfile1,outfileF,pid)
outfileC = outfileprefix +'.clusters'
# initialize cluster array
clusters = []
tcs = []
i = 0
while(i < len(fpvec1)):
print i
clusters.append(0)
tcs.append(1.0)
i = i+1
# best frist clustering.
#tc_thres = 0.4
tc_thres = tc_threshold
cluster_num = 1
for i,fp1 in enumerate(fpvec1):
print i
if (clusters[i] !=0):
continue # skip if the molecule is all ready asigned to a cluster.
clusters[i] = cluster_num # asign a cluster number to the next best molecule
for j,fp2 in enumerate(fpvec1):
if (clusters[j] !=0):
continue # skip if the molecule is all ready asigned to a cluster.
TC = tccalc.tanimoto(fp1,fp2)
#print TC
if i == j: #skip it if they are the same
continue
if (TC > tc_thres):
print cluster_num, i,j
clusters[j] = cluster_num
tcs[j] = TC
cluster_num = cluster_num + 1
file1 = open(outfileC,'w')
for i in range(len(clusters)):
file1.write('%s,%d,%f\n' % (names[i],clusters[i],tcs[i]) )
file1.close()
def main():
if (len(sys.argv) != 4): # if no input
print "ERORR"
print "syntexs: python best_frist_tc_clusters.py smiles1 outputprefix threshold"
return
smilesfile1 = sys.argv[1]
outfileprefix = sys.argv[2]
tc_threshold = float(sys.argv[3])
outfileF = outfileprefix + '.fp'
pid = str(os.getpid(
)) # get the process idenifier so that we do not right over the same file.
# read in names from smiles file.
names = []
file1 = open(smilesfile1, 'r')
for line in file1:
name = line.split()[1]
#print line, name
names.append(name)
file1.close()
# if the fp file already exists, just read in the fingerprints.
# otherwise caluclate the fingerprints.
if (os.path.isfile(outfileF)):
print outfileF + "exists."
fpvec1 = []
file1 = open(outfileF)
for line in file1:
fp = line.split()[2]
#print fp
fpvec1.append(fp)
else:
fpvec1 = tccalc.get_fp(smilesfile1, outfileF, pid)
# make dictionary of footprints.
fp_dic = {}
for i in range(len(fpvec1)):
name = names[i]
fp = fpvec1[i]
fp_dic[name] = fp
# best frist clustering.
# -- pick the top ranked molecule (frist in the list),
# -- write out everything close to that molecule,
# -- append everything else to a new list.
# -- replace the old list with the new list
name_list = names
name_newlist = []
cluster = 1
tc_thres = tc_threshold
while (len(name_list) > 0): # loop until the list is empty
print "on cluster %d" % cluster
outclusterfile = outfileprefix + ".cluster." + str(cluster)
file1 = open(outclusterfile,
'w') # open up the file for current cluster.
fp1 = fp_dic[name_list[0]] # pick the top molecule
for i, name in enumerate(name_list):
fp2 = fp_dic[name]
TC = tccalc.tanimoto(fp1, fp2)
if (TC > tc_thres):
print i
print name_list[0] + ',' + name + ":" + str(TC)
file1.write(
'%s,%f\n' % (name, TC)
) # write out everything close to that molecule, in the current cluster file.
else:
name_newlist.append(
name) # append everything else to a new list
file1.close() # close the file for the current cluster.
name_list = name_newlist # replace the old list with the new list
name_newlist = []
cluster = cluster + 1 # increment the cluster name.
def main():
if (len(sys.argv) != 4): # if no input
print "ERORR"
print "syntexs: python best_frist_tc_clusters.py smiles1 outputprefix threshold"
return
smilesfile1 = sys.argv[1]
outfileprefix = sys.argv[2]
tc_threshold = float(sys.argv[3])
outfileF = outfileprefix + '.fp'
pid = str(os.getpid(
)) # get the process idenifier so that we do not right over the same file.
# read in names from smiles file.
names = []
file1 = open(smilesfile1, 'r')
for line in file1:
name = line.split()[1]
#print line, name
names.append(name)
file1.close()
# if the fp file already exists, just read in the fingerprints.
# otherwise caluclate the fingerprints.
if (os.path.isfile(outfileF)):
print outfileF + "exists."
fpvec1 = []
file1 = open(outfileF)
for line in file1:
fp = line.split()[2]
#print fp
fpvec1.append(fp)
else:
fpvec1 = tccalc.get_fp(smilesfile1, outfileF, pid)
# make dictionary of footprints and 1-tc values.
fp_dic = {}
tc_dic = {}
for i in range(len(fpvec1)):
name = names[i]
fp = fpvec1[i]
fp_dic[name] = fp
tc_dic[name] = -1.0
# best frist clustering.
# -- pick the top ranked molecule (frist in the list),
# -- write out everything close to that molecule,
# -- append everything else to a new list.
# -- replace the old list with the new list
name_list = names
name_newlist = []
cluster = 1
tc_thres = tc_threshold
first_pass = True
while (len(name_list) > 0): # loop until the list is empty
outclusterfile = outfileprefix + ".cluster." + str(cluster)
file1 = open(outclusterfile,
'w') # open up the file for current cluster.
fp1 = fp_dic[name_list[0]] # pick the top molecule
onentc_1_2 = 1.0 - tc_dic[name_list[0]] # this is one minus the tc.
print "on cluster %d - 1-tc=%f" % (cluster, onentc_1_2)
for i, name in enumerate(name_list):
fp2 = fp_dic[name]
onentc_1_3 = 1.0 - tc_dic[name]
if ((not first_pass) and
(math.fabs(onentc_1_2 - onentc_1_3) >
(1 - tc_thres))): # by minipulating the triangle inequality.
print i, "inequality, not in cluster", onentc_1_2, onentc_1_3
name_newlist.append(name) # not put in cluster
elif ((not first_pass) and (onentc_1_2 + onentc_1_3 <=
(1 - tc_thres))):
# this should never be true here, because 2, and 3 would be already in cluster 1, if it were true, which is a contradiction.
print i, "inequality, in cluster", onentc_1_2, onentc_1_3
TC = 1 - (onentc_1_2 + onentc_1_3)
print name_list[0] + ',' + name + ": <= " + str(TC)
file1.write(
'%s,<=%f\n' % (name, TC)
) # write out everything close to that molecule, in the current cluster file.
else:
TC = tccalc.tanimoto(fp1, fp2)
if (TC > tc_thres):
print i
print name_list[0] + ',' + name + ":" + str(TC)
file1.write(
'%s,%f\n' % (name, TC)
) # write out everything close to that molecule, in the current cluster file.
else:
if (first_pass):
tc_dic[name] = TC
name_newlist.append(
name) # append everything else to a new list
first_pass = False
file1.close() # close the file for the current cluster.
name_list = name_newlist # replace the old list with the new list
name_newlist = []
cluster = cluster + 1 # increment the cluster name.