def clusteringBASIC(peaks, adds, **k):
if not peaks:
return
t=time.clock()
errorRt = k.get('rtError', 6)
#ppm = float(kwargs.get('ppm'))/10**6
ppm = k.get('ppm')
if ppm is None:
try:
ppm = peaks[0].sample.ppm/1e6
except AttributeError:
print "No value found for ppm setting to 10/1E6"
ppm = 10./1e6
#mode = k.get('mode', 'HighRes')
resolveConflicts=k.get('resolveConflicts', False)
addsToCheck=np.array(adds.keys())
adductsFound = MSPeakList()
for i, p in enumerate(peaks):
a = MSClusterList()
for v in addsToCheck:
m = p.mz+v[0]
match = peaks.peaksInMZRTRange(m, p.rt, errorRt, deltam= 2 * ppm * m)
if match is None or not match:
continue
#take the closest in mass
goodP = sorted(match, key=lambda x:abs(x.mz - (p.mz + v[0])))[0]
#if goodP in set(adductsFound):
# if resolveConflicts:
# pass
#else:
if goodP is p:
continue
a.append(goodP)
goodP.parentPeak=p
adductsFound.append(goodP)
p.fragCluster=MSPeakList(set(a))#prevent from duplicates
# def clusterComparison(list_):#receive a list of peak with clusters identified
# """
# return the best peak
# WARNING: p_ydata and p_.y_data are None
# TODO:
#
# """
# sortedList = sorted(list_, key=lambda x: len(x.fragCluster))
# longest=len(sortedList[-1].fragCluster)
# sameSizePeaks=MSPeakList()
#
# for p in sortedList:
# if len(p.fragCluster) == longest:
# sameSizePeaks.append(p)
#
# if len(sameSizePeaks) == 1:
# return sameSizePeaks[0]
# corr=np.array([0.] * len(sameSizePeaks))
# #for i, p in enumerate(sameSizePeaks):
# # for p_ in p.fragCluster:
# # corr[i] += r_coef(p_.y_data, p.y_data)
# m=max_f(corr)
# return sameSizePeaks[np.where(corr == m)[0][0]]
#
# if resolveConflicts:
# for add in set(adductsFound):
# if len(add.parentPeak) <= 1:
# #print "%s belong to several fragCluster"%str(add)
# continue
# #print "%s belong to several fragCluster"%str(add)
# goodParent=clusterComparison(add.parentPeak)
# #if goodParent is not None:
# # add.parentPeak = [goodParent]
#
# for parent in add.parentPeak:
# if parent != goodParent:
# try:
# parent.fragCluster.remove(add)
# except ValueError:
# print "Error removing %s from fragCluster of %s"%(str(add), str(parent))
# add.parentPeak = [goodParent] #the same of constructing a list 'toRemove then remove
# #print "after removing len add.parentPeak", len(add.parentPeak)
print "TiemElapsed: %s"%str(time.clock()-t)
return peaks, adductsFound
def clusteringBASIC(peaks, adds, **k):
if not peaks:
return
t = time.clock()
errorRt = k.get('rtError', 6)
#ppm = float(kwargs.get('ppm'))/10**6
ppm = k.get('ppm')
if ppm is None:
try:
ppm = peaks[0].sample.ppm / 1e6
except AttributeError:
print "No value found for ppm setting to 10/1E6"
ppm = 10. / 1e6
#mode = k.get('mode', 'HighRes')
resolveConflicts = k.get('resolveConflicts', False)
addsToCheck = np.array(adds.keys())
adductsFound = MSPeakList()
for i, p in enumerate(peaks):
a = MSClusterList()
for v in addsToCheck:
m = p.mz + v[0]
match = peaks.peaksInMZRTRange(m,
p.rt,
errorRt,
deltam=2 * ppm * m)
if match is None or not match:
continue
#take the closest in mass
goodP = sorted(match, key=lambda x: abs(x.mz - (p.mz + v[0])))[0]
#if goodP in set(adductsFound):
# if resolveConflicts:
# pass
#else:
if goodP is p:
continue
a.append(goodP)
goodP.parentPeak = p
adductsFound.append(goodP)
p.fragCluster = MSPeakList(set(a)) #prevent from duplicates
# def clusterComparison(list_):#receive a list of peak with clusters identified
# """
# return the best peak
# WARNING: p_ydata and p_.y_data are None
# TODO:
#
# """
# sortedList = sorted(list_, key=lambda x: len(x.fragCluster))
# longest=len(sortedList[-1].fragCluster)
# sameSizePeaks=MSPeakList()
#
# for p in sortedList:
# if len(p.fragCluster) == longest:
# sameSizePeaks.append(p)
#
# if len(sameSizePeaks) == 1:
# return sameSizePeaks[0]
# corr=np.array([0.] * len(sameSizePeaks))
# #for i, p in enumerate(sameSizePeaks):
# # for p_ in p.fragCluster:
# # corr[i] += r_coef(p_.y_data, p.y_data)
# m=max_f(corr)
# return sameSizePeaks[np.where(corr == m)[0][0]]
#
# if resolveConflicts:
# for add in set(adductsFound):
# if len(add.parentPeak) <= 1:
# #print "%s belong to several fragCluster"%str(add)
# continue
# #print "%s belong to several fragCluster"%str(add)
# goodParent=clusterComparison(add.parentPeak)
# #if goodParent is not None:
# # add.parentPeak = [goodParent]
#
# for parent in add.parentPeak:
# if parent != goodParent:
# try:
# parent.fragCluster.remove(add)
# except ValueError:
# print "Error removing %s from fragCluster of %s"%(str(add), str(parent))
# add.parentPeak = [goodParent] #the same of constructing a list 'toRemove then remove
# #print "after removing len add.parentPeak", len(add.parentPeak)
print "TiemElapsed: %s" % str(time.clock() - t)
return peaks, adductsFound
def isotopicPeakListFinder(peaks, isomasses, **kwargs):
"""
assign an isotopic cluster for each peak, and try to find an idms
we may use a system like the CAMERA algorithm to see...
input:
list of peak must an obj.MSPeakList object
clusterLength = 6 never go to six in LOW_RES
size expected of an isotopic cluster
rtError: maximum drift of the retention time
decreaseOrder: allow or not allow that the successive peak of the isotopic cluster
intensity are going down, can be confusing for finding idms
output:
two MSPeakList, the first one corresponding to the peaks with an isotopic cluster
and the other one all peaks belonging to an isotopic cluster
"""
#unpacking parameters
print "Isotopic cluster calculation..."
rtError = np.float(kwargs.get('rtError', 6))
ppm=np.float(peaks[0].sample.ppm/1e6)
MAX_GAP_ALLOWED = np.int(len(isomasses))
decreaseOrder = kwargs.get('decreaseOrder', True) #we use the less restrictive...
mode = kwargs.get('mode', 'Highres')
#sort isomasses
#isomasses = sorted(isomasses, key=lambda x:x[0])
peaks_with_iso =MSPeakList()
peaks_without_iso = MSPeakList()#peaks without isotopic cluster but which does not have a isotopic cluster
list_iso = set()#MSPeakList()
t = time.clock()
for peak in peaks.ipeaks():#iterating over peaks
if peak in list_iso:
continue#avoid to calculate for every peaks
isoCluster= MSClusterList()
gap = 0
#isos = resolutionAdjustment(isomasses, peak.mass()*ppm) if mode=='HighRes' else isomasses
for i, isomass in enumerate(sorted(isomasses, key=lambda x:x[0])):
#pic = _getMatchingPeaks(peaks, peak, isomass[0], ppm, rtError)
mass=isomass[0]
massToCheck=peak.mass()+mass
p = peaks.peaksInMZRange(massToCheck, deltam=ppm*massToCheck if mode=='HighRes' else 1.) #deltart
matchingRtPeaks = MSPeakList()#will contain all matching peak in rt
for pk in p.ipeaks():
if pk != peak:
if abs(peak.rt - pk.rt) <= rtError:
matchingRtPeaks.append(pk)
if matchingRtPeaks:
pic = sorted(matchingRtPeaks, key=lambda pics: abs(pics.mass()-peak.mass()))[0] #take the closest in mass
if pic is not None:
if decreaseOrder:#we want peak area inferior a peak
#if isoCluster:
areaToCompare=isoCluster[-1].area if isoCluster else peak.area
if areaToCompare < pic.area:#idms found ???
break
if pic not in list_iso:#pic not in isoCluster and
isoCluster.append(pic)
list_iso.add(pic)
else:
gap+=1
if gap >=MAX_GAP_ALLOWED:
break
# #set parent for all peaks found
if isoCluster:
for pics in isoCluster:
#pics.parentPeak=peak
pics.parentPeak.append(peak)
peak.isoCluster = isoCluster
peaks_with_iso.addPeak(peak)
else:
peaks_without_iso.addPeak(peak)
# for p in peaks.ipeaks():
# if p not in peaks_with_iso and p not in list_iso:
# peaks_without_iso.addPeak(p)
print time.clock()-t
print "peaks with isotopes: " ,len(peaks_with_iso)
print "list isotopes: " ,len(list_iso)
print "peaks without isotopes: " ,len(peaks_without_iso)
return peaks_with_iso+peaks_without_iso, list_iso
def isotopicPeakListFinder(peaks, isomasses, **kwargs):
"""
assign an isotopic cluster for each peak, and try to find an idms
we may use a system like the CAMERA algorithm to see...
input:
list of peak must an obj.MSPeakList object
clusterLength = 6 never go to six in LOW_RES
size expected of an isotopic cluster
rtError: maximum drift of the retention time
decreaseOrder: allow or not allow that the successive peak of the isotopic cluster
intensity are going down, can be confusing for finding idms
output:
two MSPeakList, the first one corresponding to the peaks with an isotopic cluster
and the other one all peaks belonging to an isotopic cluster
"""
#unpacking parameters
print "Isotopic cluster calculation..."
rtError = np.float(kwargs.get('rtError', 6))
ppm = np.float(peaks[0].sample.ppm / 1e6)
MAX_GAP_ALLOWED = np.int(len(isomasses))
decreaseOrder = kwargs.get('decreaseOrder',
True) #we use the less restrictive...
mode = kwargs.get('mode', 'Highres')
#sort isomasses
#isomasses = sorted(isomasses, key=lambda x:x[0])
peaks_with_iso = MSPeakList()
peaks_without_iso = MSPeakList(
) #peaks without isotopic cluster but which does not have a isotopic cluster
list_iso = set() #MSPeakList()
t = time.clock()
for peak in peaks.ipeaks(): #iterating over peaks
if peak in list_iso:
continue #avoid to calculate for every peaks
isoCluster = MSClusterList()
gap = 0
#isos = resolutionAdjustment(isomasses, peak.mass()*ppm) if mode=='HighRes' else isomasses
for i, isomass in enumerate(sorted(isomasses, key=lambda x: x[0])):
#pic = _getMatchingPeaks(peaks, peak, isomass[0], ppm, rtError)
mass = isomass[0]
massToCheck = peak.mass() + mass
p = peaks.peaksInMZRange(
massToCheck,
deltam=ppm *
massToCheck if mode == 'HighRes' else 1.) #deltart
matchingRtPeaks = MSPeakList(
) #will contain all matching peak in rt
for pk in p.ipeaks():
if pk != peak:
if abs(peak.rt - pk.rt) <= rtError:
matchingRtPeaks.append(pk)
if matchingRtPeaks:
pic = sorted(matchingRtPeaks,
key=lambda pics: abs(pics.mass() - peak.mass()))[
0] #take the closest in mass
if pic is not None:
if decreaseOrder: #we want peak area inferior a peak
#if isoCluster:
areaToCompare = isoCluster[
-1].area if isoCluster else peak.area
if areaToCompare < pic.area: #idms found ???
break
if pic not in list_iso: #pic not in isoCluster and
isoCluster.append(pic)
list_iso.add(pic)
else:
gap += 1
if gap >= MAX_GAP_ALLOWED:
break
# #set parent for all peaks found
if isoCluster:
for pics in isoCluster:
#pics.parentPeak=peak
pics.parentPeak.append(peak)
peak.isoCluster = isoCluster
peaks_with_iso.addPeak(peak)
else:
peaks_without_iso.addPeak(peak)
# for p in peaks.ipeaks():
# if p not in peaks_with_iso and p not in list_iso:
# peaks_without_iso.addPeak(p)
print time.clock() - t
print "peaks with isotopes: ", len(peaks_with_iso)
print "list isotopes: ", len(list_iso)
print "peaks without isotopes: ", len(peaks_without_iso)
return peaks_with_iso + peaks_without_iso, list_iso
