class damageSeriesAnalysis():
# For damage series from PDB run the ETRACK scripts to calculate per-atom density metrics.
# Below are a series of methods for processing specific damage series within the pdb.
def __init__(self):
self.ETRACK = ETRACK()
def runDataseries(self,name,version,process,postprocess,retrieve):
if name == 'TRAP':
self.writeETRACKinputfile_TRAP(version)
elif name == 'BURM':
self.writeETRACKinputfile_Burm2000(version)
elif name == 'FIOR':
self.writeETRACKinputfile_Fior2007()
elif name == 'DELAMORA':
self.writeETRACKinputfile_DelaMora2011(version)
elif name == 'FRANK':
self.writeETRACKinputfile_Frankaer2014()
elif name == 'JUERS100':
self.writeETRACKinputfile_Juers2011_100K()
elif name == 'JUERS160':
self.writeETRACKinputfile_Juers2011_160K()
elif name == 'BURY':
self.writeETRACKinputfile_Bury2015(version)
elif name == 'WEIK':
self.writeETRACKinputfile_Weik2000(version)
elif name == 'DIXON':
self.writeETRACKinputfile_TDIXONinsulin(version)
elif name == 'SUTTON':
self.writeETRACKinputfile_Sutton2013()
elif name == 'PETROVA':
self.writeETRACKinputfile_Petrova2010(version)
elif name == 'NANAO':
self.writeETRACKinputfile_Nanao2005(version)
else:
'Data series not recognised'
return
self.ETRACK.runETRACK(process,postprocess,retrieve)
def getDensSeries(self):
dSeries = [['DELAMORA','DIFF'],
['JUERS100',''],
['JUERS160',''],
['BURY','initial'],
['WEIK','initial'],
['DIXON','DIFF'],
['FIOR',''],
['BURM','final'],
['TRAP','DIFF'],
['SUTTON',''],
['PETROVA','DIFF'],
['NANAO','Elastase'],
['NANAO','Thaumatin'],
['NANAO','Trypsin'],
['NANAO','Lysozyme'],
['NANAO','Insulin'],
['NANAO','RibonucleaseA']]
return dSeries
def writeETRACKinputfile_TRAP(self,version):
# Need to create input file for this test TRAP damage series
doses = '1.31,3.88,6.45,9.02,11.58,14.15,16.72,19.29,21.86,24.98'
dNames = '1,2,3,4,5,6,7,8,9,10'
doses,dNames = self.ETRACK.defineDoseList(doses,dNames,version)
inputString = self.ETRACK.writeInputFile('/Users/charlie/DPhil/YEAR2/JAN/TRAP_ETRACK/{}/'.format(version),
'TRAP',dNames,'TRAP1.pdb',doses,'TRAP_data.pkl')
def writeETRACKinputfile_Burm2000(self,version):
# Need to create input file for the Burmeister 2000 damage series
# version in ('final','initial') depending on which PDB_REDO version required
doses = '1,2,3,4'
dNames = 'a,f,g,h'
doses,dNames = self.ETRACK.defineDoseList(doses,dNames,'DIFF')
inputString = self.ETRACK.writeInputFile('/Users/charlie/DPhil/YEAR2/JAN/Burm2000_ETRACK/{}/'.format(version),
'1dw',dNames,'1dwa.pdb',doses,'1dw_data.pkl')
def writeETRACKinputfile_Sutton2013(self):
# Need to create input file for the Sutton 2013 damage series
doses = '1,2,3,4,5,6,7,8,9,10,11,12,13,14,15'
dNames = '8x,8y,8z,9a,9b,9c,9e,9f,9h,9i,90,91,92,93,94'
doses,dNames = self.ETRACK.defineDoseList(doses,dNames,'DIFF')
inputString = self.ETRACK.writeInputFile('/Users/charlie/DPhil/YEAR2/JAN/Sutton2013_ETRACK/DIFF/',
'4h',dNames,'4h8x.pdb',doses,'4h_data.pkl')
def writeETRACKinputfile_Petrova2010(self,version):
# Need to create input file for the Petrova 2010 damage series
dNames = '1,2,3,4,5,6,7,8'
doses = '1.2,14.2,15.4,28.4,29.6,42.6,43.8,56.8'
doses,dNames = self.ETRACK.defineDoseList(doses,dNames,version)
inputString = self.ETRACK.writeInputFile('/Users/charlie/DPhil/YEAR2/JAN/Petrova2010_ETRACK/M_100K/{}/'.format(version),
'PET',dNames,'PET1.pdb',doses,'PET_data.pkl')
def writeETRACKinputfile_Nanao2005(self,protein):
# Need to create input file for the Nanao 2005 damage sets
# 'protein' in ('Elastase','Thaumatin','Trypsin','Lysozyme','Insulin','RibonucleaseA')
pInfo = {'Trypsin':'lv,lw','Thaumatin':'lr,lu','Elastase':'lo,lq',
'Lysozyme':'lx,ly','RibonucleaseA':'lp,lz','Insulin':'n3,n1'}
if protein not in pInfo.keys():
return
doses = '1,2'
dNames = pInfo[protein]
doses,dNames = self.ETRACK.defineDoseList(doses,dNames,'DIFF')
inputString = self.ETRACK.writeInputFile('/Users/charlie/DPhil/YEAR2/JAN/Nanao2005_ETRACK/{}/DIFF/'.format(protein),
'2b',dNames,'2b{}.pdb'.format(pInfo[protein].split(',')[0]),doses,'2b_data.pkl')
def writeETRACKinputfile_DelaMora2011(self,version):
# Need to create input file for the DelaMora 2011 damage series
# type in ('DIFF','SIMPLE','END')
doses = '2.31,6.62,12.31,17.9,23.3,28.6'
dNames = 'h,i,j,l,m,n'
doses,dNames = self.ETRACK.defineDoseList(doses,dNames,version)
inputString = self.ETRACK.writeInputFile('/Users/charlie/DPhil/YEAR2/JAN/DelaMora2011_ETRACK/{}/'.format(version),
'2yb',dNames,'2ybh.pdb',doses,'2yb_data.pkl')
def writeETRACKinputfile_Fior2007(self):
# Need to create input file for the Fioravanti 2007 series
doses = '1,2,3'
dNames = 'k,q,r'
doses,dNames = self.ETRACK.defineDoseList(doses,dNames,'DIFF')
inputString = self.ETRACK.writeInputFile('/Users/charlie/DPhil/YEAR2/JAN/Fioravanti2007_ETRACK/',
'2j5',dNames,'2j5k.pdb',doses,'2j5_data.pkl')
def writeETRACKinputfile_Frankaer2014(self):
# Need to create input file for the Frankaer 2014 series
doses = '2,3,4'
dNames = 'h,i,j'
doses,dNames = self.ETRACK.defineDoseList(doses,dNames,'DIFF')
inputString = self.ETRACK.writeInputFile('/Users/charlie/DPhil/YEAR2/JAN/Frankaer2014_ETRACK/',
'4m4',dNames,'4m4f.pdb',doses,'4m4_data.pkl')
def writeETRACKinputfile_Juers2011_100K(self):
# Need to create input file for the Frankaer 2014 100K series
doses = '1,2,3,4'
dNames = 'p,q,r,s'
doses,dNames = self.ETRACK.defineDoseList(doses,dNames,'DIFF')
inputString = self.ETRACK.writeInputFile('/Users/charlie/DPhil/YEAR2/JAN/Juers2011-100K_ETRACK/',
'3p7',dNames,'3p7p.pdb',doses,'3p7_data.pkl')
def writeETRACKinputfile_Juers2011_160K(self):
# Need to create input file for the Frankaer 2014 160K series
doses = '1,2,3,4'
dNames = 't,u,v,w'
doses,dNames = self.ETRACK.defineDoseList(doses,dNames,'DIFF')
inputString = self.ETRACK.writeInputFile('/Users/charlie/DPhil/YEAR2/JAN/Juers2011-160K_ETRACK/',
'3p7',dNames,'3p7t.pdb',doses,'3p7_data.pkl')
def writeETRACKinputfile_Bury2015(self,version):
# Need to create input file for the Bury2015 series
# version in ('final','initial') depending on which PDB_REDO version required
doses = '2.1,6.2,10.3,14.4,20.6,26.8,35.7,44.6'
dNames = 'b,c,d,e,f,g,h,i'
doses,dNames = self.ETRACK.defineDoseList(doses,dNames,'DIFF')
inputString = self.ETRACK.writeInputFile('/Users/charlie/DPhil/YEAR2/JAN/Bury2015_ETRACK/{}/'.format(version),
'4x4',dNames,'4x4b.pdb',doses,'4x4_data.pkl')
def writeETRACKinputfile_Weik2000(self,version):
# Need to create input file for the Weik2000 series
# version in ('final','initial') depending on which PDB_REDO version required
doses = '1,2,3,4,5,6,7,8,9'
dNames = 'd,e,f,g,h,i,j,k,m'
doses,dNames = self.ETRACK.defineDoseList(doses,dNames,'DIFF')
inputString = self.ETRACK.writeInputFile('/Users/charlie/DPhil/YEAR2/JAN/Weik2000_ETRACK/{}/'.format(version),
'1qi',dNames,'1qid.pdb',doses,'1qi_data.pkl')
def writeETRACKinputfile_TDIXONinsulin(self,version):
# Need to create input file for the TDIXON-insulin series
doses = '0.89,2.74,4.59,6.44,8.29,10.14,11.98,13.84,15.68,17.54'
dNames = '1,2,3,4,5,6,7,8,9,10'
doses,dNames = self.ETRACK.defineDoseList(doses,dNames,version)
inputString = self.ETRACK.writeInputFile('/Users/charlie/DPhil/YEAR2/JAN/TDIXON_InsulinSeries_ETRACK/{}/'.format(version),
'insu',dNames,'insu1.pdb',doses,'insu_data.pkl')
def writeETRACKinputfile_TDIXONinsulin_FEB(self,version):
# Need to create input file for the TDIXON-insulin series
doses = '0.89,2.74,4.59,6.44,8.29,10.14,11.98,13.84,15.68,17.54'
dNames = '1,2,3,4,5,6,7,8,9,10'
doses,dNames = self.ETRACK.defineDoseList(doses,dNames,version)
inputString = self.ETRACK.writeInputFile('/Users/charlie/DPhil/YEAR2/FEB/ETRACK-testing/2FOFC-testing/TDinsulin/',
'insu',dNames,'insu1.pdb',doses,'insu_data.pkl')
def residueMetricDistributionPlots(self,mapPro,postPro,retr,densMet,normType):
# for each damage series retrieve the distribution for damage metric values for all atoms of
# specified residues (see 'resList' below), and then plot a kde plot for each
resList = [['GLU','GLN'],['ASP','ASN'],['ILE','LEU'],['TYR','PHE'],
['TYR','ASP','GLU'],['TYR','PHE','GLY'],
['GLU','GLY'],['ASP','GLY'],['TYR','GLY'],['PHE','GLY'],
['CYS','GLY'],['MET','GLY'],['GLU','ASP','CYS','MET','TYR']]
dSeries = self.getDensSeries()
plotData = {'-'.join(res):{} for res in resList}
for key in plotData.keys():
for k in key.split('-'):
plotData[key][k] = []
for dSer in dSeries:
self.runDataseries(dSer[0],dSer[1],mapPro,postPro,retr)
for resGroup in resList:
if len(resGroup) > 4:
plotType = 'kde'
else:
plotType = 'both'
data = self.ETRACK.et.combinedAtoms.graphMetricDistn(densMet,normType,True,plotType,resGroup,True)
for r in resGroup:
plotData['-'.join(resGroup)][r] += data[r]
for resGroup in resList:
# sns.set_palette("deep", desat=.6)
sns.set_style("whitegrid")
sns.set_context(rc={"figure.figsize": (10, 6)})
fig = plt.figure()
ax = plt.subplot(111)
if len(resGroup) > 4:
plotType = 'kde'
else:
plotType = 'both'
for j,(res, color) in enumerate(zip(resGroup, sns.color_palette('hls', len(resGroup)))):
datax = plotData['-'.join(resGroup)][res]
print 'number of {} atoms = {}'.format(res,len(datax))
self.ETRACK.et.combinedAtoms.plotHist(plotType=plotType,datax=datax,
lbl='average,{}'.format(res),color=color)
plt.legend()
plt.xlabel('{} D{} per atom'.format(normType,densMet),fontsize=18)
plt.ylabel('Norm-frequency',fontsize=18)
plt.title('{} D{} per atom, residues: {}'.format(normType,densMet,','.format(resGroup)))
saveName = '{}_{}D{}-{}-COMBINED.png'.format(''.join(resGroup),normType.replace(" ",""),densMet,plotType)
fig.savefig(saveName)
def findTyrDlossVNeighbourhood(self,mapPro,postPro,retr,densMet,normType,resType,atomType,distance,weighted):
# for each damage series determine the speicified damage metric for all resType-atomType atoms and compare
# to the average density metric within a local distance of each atom.
# 'distance' is distance from an atom in Angstrom
# 'weighted' is True if distance-weighted average metric taken over local environment, otherwise,
# if False then standard average of metric taken.
combPlotData = {'x':[],'y':[]}
dSeries = self.getDensSeries()
for dSer in dSeries:
self.runDataseries(dSer[0],dSer[1],mapPro,postPro,retr)
if normType == 'Calpha normalised':
self.et.combinedAtoms.calcAdditionalMetrics(densMet,normType,'Calpha')
self.ETRACK.et.combinedAtoms.calcAdditionalMetrics(densMet,normType,'average')
plotData = self.ETRACK.et.combinedAtoms.calculateLocalDloss(resType,atomType,distance,densMet,normType,weighted)
for k in plotData.keys():
combPlotData[k] += plotData[k]
# plot the relationship between atom Dloss and local environment average Dloss
self.ETRACK.et.combinedAtoms.plotScatterPlot(combPlotData['x'],combPlotData['y'],
'{}-{} D{}'.format(resType,atomType,densMet),
'local environment D{}'.format(densMet),
'Average D{} within {} Angstrom of {}-{}'.format(densMet,distance,resType,atomType),
'D{}Scatter_{}-{}_localEnvironmentComparison-COMBINED.png'.format(densMet,resType,atomType),
True,False,'')
def rankTyrOHdamage(self,mapPro,postPro,retr,densMet,normType):
# for each damage series determine the specified damage metric for all TYR-OH atoms and rank them
dSeries = self.getDensSeries()
tyrOHatms = {'atmID':[],'metric':[]}
for dSer in dSeries:
self.runDataseries(dSer[0],dSer[1],mapPro,postPro,retr)
if normType == 'Calpha normalised':
self.et.combinedAtoms.calcAdditionalMetrics(densMet,normType,'Calpha')
self.ETRACK.et.combinedAtoms.calcAdditionalMetrics(densMet,normType,'average')
atms = self.ETRACK.et.combinedAtoms.getAtom('','TYR','','OH')
for a in atms:
tyrOHatms['atmID'].append(dSer[0]+'-'+dSer[1]+'-'+a.getAtomID())
tyrOHatms['metric'].append(a.densMetric[densMet][normType]['average'])
# sort the Tyr-OH atms in order of metric value:
list1, list2 = (list(t) for t in zip(*sorted(zip(tyrOHatms['metric'], tyrOHatms['atmID']))))
for i in range(len(list1)):
print '{} ---> {}'.format(list2[i],list1[i])
def runBatchSeries(self,mapPro,postPro,retr,densMet,normType):
# plot scatter plots of damage to GLU/ASP/TYR side-chain atoms to compare the relative damage to adjacent atoms
# e.g. TYR-OH vs TYR-CZ or GLU-CD vs GLU-CG etc.
dSeries = self.getDensSeries()
rSquaredDic,numPairsDic,dataDic = {},{},{}
for dSer in dSeries:
self.runDataseries(dSer[0],dSer[1],mapPro,postPro,retr)
# calculate additional metrics as required
self.ETRACK.et.combinedAtoms.calcMetricDiffFromStructureMean(densMet,'Standard','std-devs')
if normType == 'Calpha normalised':
self.ETRACK.et.combinedAtoms.calcAdditionalMetrics(densMet,normType,'Calpha')
rSquared,numPairs,data = self.ETRACK.et.combinedAtoms.compareSensAtoms(densMet,normType)
rSquaredDic[''.join(dSer)] = rSquared
numPairsDic[''.join(dSer)] = numPairs
for key in data.keys():
if key not in dataDic.keys(): dataDic[key] = {'x':[],'y':[]}
for i in ('x','y'):
dataDic[key][i] += data[key][i]#+= list((np.array(data[key][i])-avMetric*np.ones(len(data[key][i])))/stdMetric)
# plot this data as scatter plot
for key in dataDic.keys():
xData = dataDic[key]['x']
yData = dataDic[key]['y']
k = key.split('_')
xLabel = '{}-{} D{}'.format(k[0],k[1],densMet)
yLabel = '{}-{} D{}'.format(k[0],k[2],densMet)
rSquared = self.ETRACK.et.combinedAtoms.plotScatterPlot(xData,yData,xLabel,
yLabel,'D{} for {} atoms'.format(densMet,key),
'D{}_{}_scatterplotCOMBINED.svg'.format(densMet,key),True,True,'')
return rSquaredDic,numPairsDic,dataDic
def runBatchSeries2(self,mapPro,postPro,retr,densMet,normType,rType):
pairs = [['GLU','CD','CG'],['GLU','CD','OE1'],['ASP','CG','CB'],
['ASP','CG','OD1'],['TYR','OH','CZ']]
dSeries = self.getDensSeries()
seriesData = {}
for dSer in dSeries:
self.runDataseries(dSer[0],dSer[1],mapPro,postPro,retr)
self.ETRACK.et.combinedAtoms.calcMetricDiffFromStructureMean(densMet,'Standard','std-devs')
data = self.ETRACK.et.combinedAtoms.findMetricRatioKeyResidues_scatterplot(densMet,normType,rType,pairs,dSer)
seriesData['-'.join(dSer)] = data
# plot a combined dataseries plot here
sns.set_palette("deep", desat=.6)
sns.set_context(rc={"figure.figsize":(40, 10)})
fig = plt.figure()
ax = plt.subplot(111)
colors = ['#737474','#409cd6','#58bb6b','#faa71a','#ff6a6a']
i = -2
for key in seriesData.keys():
i += 2
for pair in seriesData[key].keys():
j = (['-'.join(p) for p in pairs]).index(pair)
x = seriesData[key][pair]['x']
y = seriesData[key][pair]['y']
xNew = np.array(x)+i
if i == 0:
plt.scatter(xNew,y,marker='o',s=100,c=colors[j],edgecolors='#FFFFFF',label=pair)
else:
plt.scatter(xNew,y,marker='o',s=100,c=colors[j],edgecolors='#FFFFFF')
plt.plot([-1, i+2],[0,0],':',color='#088DA5') # a horizontal line at y=0
tickPts = list(2*np.array(range(len(seriesData.keys())))+0.5)
ax.set_xlim([-1, i+2])
plt.xticks(tickPts,seriesData.keys())
plt.xlabel('Damage Series', fontsize=24)
plt.ylabel('D{} {}'.format(densMet,rType), fontsize=24)
figtitle = '{} D{} {}'.format(normType,densMet,rType)
# place legend outside to right of plot
box = ax.get_position()
ax.set_position([box.x0, box.y0, box.width * 0.95, box.height])
ax.legend(loc='center left', bbox_to_anchor=(1, 0.5),fontsize=24)
fig.suptitle(figtitle,fontsize=28)
saveTitle = figtitle.replace(' ','_')
fname = lambda x: saveTitle.strip('.png')+'_{}.png'.format(x)
i = 0
while os.path.isfile(fname(i)): i += 1
fig.savefig(fname(i))
def runBatchSeries3(self,mapPro,postPro,retr,densMet,normType,distLim):
seriesData = {}
dSeries = self.getDensSeries()
plotData1 = {'x':[],'y':[],'colors':[]}
plotData2 = {'x':[],'y':[]}
for dSer in dSeries:
self.runDataseries(dSer[0],dSer[1],mapPro,postPro,retr)
for diff in ('std-devs','ratio'):
self.ETRACK.et.combinedAtoms.calcMetricDiffFromStructureMean(densMet,'Standard',diff)
if normType == 'Calpha normalised':
self.ETRACK.et.combinedAtoms.calcAdditionalMetrics(densMet,normType,'Calpha')
self.ETRACK.et.combinedAtoms.calcAdditionalMetrics('loss',normType,'average')
self.ETRACK.et.combinedAtoms.seriesName = '-'.join(dSer)
# get pdb file information and space group
pdbName = self.ETRACK.et.where+self.et.initialPDB
sGroup = self.ETRACK.et.getSpaceGroup()
if sGroup is False: return
# determine correlation between TYR-OH damage and carboxyl contacts
seriesData['-'.join(dSer)],TyrDam,CO2Dam,scatterColor = self.ETRACK.et.combinedAtoms.densMetSurroundAtmsCorrel('TYR','OH',distLim,normType,densMet,
True,pdbName,sGroup)
# data for Glu-CD/Asp-CG Dloss vs Tyr-OH Dloss
plotData1['x'] += TyrDam
plotData1['y'] += CO2Dam
plotData1['colors'] += scatterColor
# determine the per-dataset correlation between solvent accessibility and Tyr-OH Dloss
solvAccDic,plotData = self.ETRACK.et.combinedAtoms.compareSolvAccessWithAndWithoutGluAspGroups(pdbName,sGroup,'TYR','OH',[],densMet,normType)
plotData2['x'] += plotData['x']
plotData2['y'] += plotData['y']
for key in seriesData.keys():
print 'For dataseries: {}'.format(key)
print seriesData[key]
# plot the relationship between TYR-OH density and nearby carboxyl atom density
self.ETRACK.et.combinedAtoms.plotScatterPlot(plotData1['x'],
plotData1['y'],
'TYR-OH D{}'.format(densMet),
'Carboxyl-contact D{}'.format(densMet),
'TYR-OH vs carboxyl-contact D{}'.format(densMet),
'D{}Scatter_TYR-OH_carboxylContacts-COMBINED.png'.format(densMet),
True,
False,
plotData1['colors'])
# plot the relationship between TYR-OH density and solvent accessibility
self.ETRACK.et.combinedAtoms.plotScatterPlot(plotData2['x'],
plotData2['y'],
'Solvent Accessibility',
'TYR-OH D{}'.format(densMet),
'TYR-OH D{} vs solvent accessibility'.format(densMet),
'{}D{}Scatter_TYR-OH_solvAccess-COMBINED.png'.format(normType,densMet),
True,False,
'')
def __init__(self): self.ETRACK = ETRACK()