def _fitLinear(self): '''Private method: Fits the linear segments of the data''' if self.linearFitted is False: expdata = zip(self.foldedRegion['temperatures'], self.foldedRegion['data']) d,f = Fitting.doFit(model='Linear', conv=1E-11, expdata=expdata, silent=True) self.foldedSlope = d[0] self.foldedIntercept = d[1] #The value of the heat capacity at the start of the transition self.foldedOffset = self.foldedSlope*self.temperatures[self.foldedIndexes[1]] + self.foldedIntercept if self.useFolded is True or (len(self.unfoldedRegion['temperatures']) < 2): self.unfoldedSlope = self.foldedSlope unfoldedHC = self.data[self.unfoldedIndexes[0]] foldedHC = self.foldedSlope*self.temperatures[self.unfoldedIndexes[0]] + self.foldedIntercept self.unfoldedIntercept = self.foldedIntercept + unfoldedHC - foldedHC #print 'Unfolded intercept derived from Folded ', self.unfoldedIntercept else: expdata = zip(self.unfoldedRegion['temperatures'], self.unfoldedRegion['data']) d,f = Fitting.doFit(model='Linear', conv=1E-11, expdata=expdata, silent=True) self.unfoldedSlope = d[0] self.unfoldedIntercept = d[1] #print 'Unfolded intercept derived from fit ',self.unfoldedIntercept #Calculate the heat capacity difference (DeltaCp) at the start of the unfolded region unfoldedHC = self.data[self.unfoldedIndexes[0]] foldedHC = self.foldedSlope*self.temperatures[self.unfoldedIndexes[0]] + self.foldedIntercept self.deltaHC = unfoldedHC - foldedHC self.linearFitted = True
def _fitLinear(self):
'''Private method: Fits the linear segments of the data'''
if self.linearFitted is False:
expdata = zip(self.foldedRegion['temperatures'],
self.foldedRegion['data'])
d, f = Fitting.doFit(model='Linear',
conv=1E-11,
expdata=expdata,
silent=True)
self.foldedSlope = d[0]
self.foldedIntercept = d[1]
#The value of the heat capacity at the start of the transition
self.foldedOffset = self.foldedSlope * self.temperatures[
self.foldedIndexes[1]] + self.foldedIntercept
if self.useFolded is True or (len(
self.unfoldedRegion['temperatures']) < 2):
self.unfoldedSlope = self.foldedSlope
unfoldedHC = self.data[self.unfoldedIndexes[0]]
foldedHC = self.foldedSlope * self.temperatures[
self.unfoldedIndexes[0]] + self.foldedIntercept
self.unfoldedIntercept = self.foldedIntercept + unfoldedHC - foldedHC
#print 'Unfolded intercept derived from Folded ', self.unfoldedIntercept
else:
expdata = zip(self.unfoldedRegion['temperatures'],
self.unfoldedRegion['data'])
d, f = Fitting.doFit(model='Linear',
conv=1E-11,
expdata=expdata,
silent=True)
self.unfoldedSlope = d[0]
self.unfoldedIntercept = d[1]
#print 'Unfolded intercept derived from fit ',self.unfoldedIntercept
#Calculate the heat capacity difference (DeltaCp) at the start of the unfolded region
unfoldedHC = self.data[self.unfoldedIndexes[0]]
foldedHC = self.foldedSlope * self.temperatures[
self.unfoldedIndexes[0]] + self.foldedIntercept
self.deltaHC = unfoldedHC - foldedHC
self.linearFitted = True
def do_search(self, globalop, proteinop, proteins, datasets):
"""Do searches for the various parameters, name, pka etc"""
import re, urllib
from PEATDB.PEATTables import PEATTableModel
from PEATDB.Ekin.Fitting import Fitting
from PEATDB.Ekin.Web import EkinWeb
DB = self.DB
EW = EkinWeb()
ekincols = DB.ekintypes
found=[]
protlist = proteins.split(' ')
residuelist = residues.split(' ')
def createPhrase(items, op):
if op == 'or':
logicsymbol = '|'
else:
logicsymbol = '+'
itemlist = items.split(' ')
phrase =''
c=1
for i in itemlist:
if c == len(itemlist):
phrase = phrase + i
else:
phrase = phrase + i + logicsymbol
c=c+1
return re.compile(phrase, re.IGNORECASE)
#if there is a protein name entered, get the list of proteins/records to use
#otherwise use all records to search for the other keys
names = []
keywords = {}
for p in DB.getRecs():
name = DB[p].name
names.append((name,p))
if hasattr(DB[p], 'keywords'):
keywords[name] = DB[p].keywords
else:
keywords[name] = ''
names.sort()
#create search expressions
s_protein = createPhrase(proteins, proteinop)
s_residue = createPhrase(residues, 'or')
pkarange = pka.split('-')
#do search
for name in names:
proteinname = name[0]
protein = name[1]
if s_protein.search(proteinname) or s_protein.search(keywords[proteinname]):
found.append(protein)
if len(found)==0:
print '<h2>Sorry, no proteins found that match this name. <h2>'
return
elif residues == '' and pka == '':
self.show_DB(selected=found)
return found
#now search for the other keys inside selected proteins if needed
ekinfound={}; foundfits={}
ignore =['__fit_matches__', '__Exp_Meta_Dat__', '__datatabs_fits__']
kys = list(DB.getRecs())
kys.sort()
if len(found) == 0 or globalop == 'or':
found = DB.getRecs()
print '<table id="mytable" cellspacing=0 align=center>'
'''search recs for residue and pkarange
and add dataset names to new ekinfound dict'''
for protein in found:
for col in DB[protein].keys():
if nucleus != 'any':
if nucleus not in col:
continue
E = DB[protein][col]
if DB['userfields'].has_key(col):
fieldtype=DB['userfields'][col]['field_type']
else:
fieldtype=None
if fieldtype in ekincols:
dk = E.datasets
fits = E.__datatabs_fits__
for k in dk:
meta = E.getMetaData(k)
try:
thisres = meta['residue']
except:
thisres = k
if thisres == None:
thisres = k
if residues != '':
if s_residue.search(thisres) and not k in ignore:
if not ekinfound.has_key(protein+'$'+col):
ekinfound[protein+'$'+col]=[]
ekinfound[protein+'$'+col].append(k)
if pka != '':
foundpka = 0
if k in fits.keys():
if fits[k] == None:
continue
if fits[k].has_key('model'):
model = fits[k]['model']
else:
continue
if not foundfits.has_key(protein+'$'+col):
foundfits[protein+'$'+col]={}
X = Fitting.getFitter(model)
pnames = X.names
i=0
#check if first num is larger, then swap
try:
pk1=float(pkarange[0])
pk2=float(pkarange[1])
except:
print '<h2> Error: pka values are not valid </h2>'
return
if pk1 > pk2:
tmp = pk1
pk1 = pk2
pk2 = tmp
#iterate thru parameter names and match any pK fields
#this code is not that efficient!
for p in pnames:
if 'pK' in p:
pka = fits[k][i]
if pka >= pk1 and pka <= pk2:
foundpka=1
i=i+1
#if match is 'ANY', just append dataset if not there already
#also for case if no residue value entered
if globalop == 'or' or residues == '':
if foundpka == 1:
if not ekinfound.has_key(protein+'$'+col):
ekinfound[protein+'$'+col]=[]
if not k in ekinfound[protein+'$'+col]:
ekinfound[protein+'$'+col].append(k)
#if match is 'ALL', need to check dataset already found
#and if no pka found, remove it. if both are true keep it
elif globalop == 'and':
if foundpka == 0:
if ekinfound.has_key(protein+'$'+col) and k in ekinfound[protein+'$'+col]:
#print 'removing', protein, col
ekinfound[protein+'$'+col].remove(k)
foundfits[protein+'$'+col][k]=fits[k]
#check for empty fields in ekinfound dict and delete them..
for d in ekinfound.keys():
if len(ekinfound[d])==0:
del(ekinfound[d])
#if no results, just say that and return
if len(ekinfound)==0:
print '<h2> Sorry, no records found that match these parameters. </h2>'
return
#top display button and options
print '<form name="resultsform" action="%s/main.cgi" METHOD="POST" ENCTYPE="multipart/form-data">' %self.bindir
self.write_sessionkey('show_datasets')
print '<td valign=top align=right colspan=3> <input type=submit value="display selected" name=submit>'
print '<label><input type=checkbox name="plotoption" value=3>single graph</label>'
print '<label><input type=checkbox name="normalise" value=1>normalise</label>'
print '<label><input type=checkbox name="logx" value=1>log-x</label>'
print '<label><input type=checkbox name="logy" value=1>log-y</label>'
print '<label><input type=checkbox name="legend" value=1>legend</label>'
print '<input type=button value="Check All" onClick="checkAll(document.resultsform.residue)">'
print '<input type=button value="Uncheck All" onClick="uncheckAll(document.resultsform.residue)"></td>'
print '</tr>'
#header
cols=['protein','column','residues']
for c in cols:
print '<th>'+c+'</th>'
print '</tr>'
ekys = ekinfound.keys()
ekys.sort()
r=1
for k in ekys:
if r % 2 == 0:
cls = "spec"
else:
cls = ""
fields = k.split('$')
protein = fields[0]
column = fields[1]
proteinname = DB[protein]['name']
print '<tr>'
print '<th class="spec">%s</th> <td> %s </td>' %(proteinname, column)
print '<td>'
for residue in ekinfound[k]:
residuefield = k+"$"+residue
fithtml = ''
try:
print '<input type=checkbox id="residue" name="residue" value=%s>' %("'"+residuefield+"'")
except:
print 'UnicodeEncodeError'
continue
urlfields = urllib.urlencode({'login':self.user,'project':self.project,
'residue': residuefield,'action': 'show_datasets'})
print '<a href="/cgi-bin/titration_db/main.cgi?%s" target="_blank">%s</a>'\
% (urlfields, residue)
print '</tr>'
r=r+1
print '</form>'
print '</table>'
self.footer()
return
def fit(self, model): '''Fits the normalised curve to the specified dsc model Params: model - A string. Valid values are TwoState, NonTwoState and Irreversible. Invalid values cause a ValueError exception to be raised Returns: A tuple. The first element is a list containing the fit data. Strings decribing each element can be obtained by calling fitHeaders(model). The second element is the fit error''' if len(self.unfoldedRegion['temperatures']) < 2 or self.useFolded is True: method = 'FoldedOnly' else: method = 'Both' dscCurve = self.normalisedCurve() #conver to kj - fitting module has R in kj value = [el*4.184 for el in dscCurve.value] #Use area under curve (calculated in progressBaseline) as enthalpy guess #Melting temp guess is the temperature with max cp enthalpyGuess = self.curveArea*4.184 meltingGuess = dscCurve.temperature[value.index(max(value))] if model == 'TwoState': #Parameters Tm, VantHoff startValues = [meltingGuess, enthalpyGuess] fittingParameters,fittingInstance=Fitting.doFit(expdata=zip(dscCurve.temperature,value), model='DSC2state', guess=False, startvalues=startValues, silent=True) #Use fitData since it gives dict entries names - increased readibility fitData = fittingInstance.getResult() vantHoff = fitData['deltaH']/4.184 meltingTemp = fitData['Tm'] res = ['TwoState', method, self.foldedRange[1], self.unfoldedRange[0], meltingTemp, vantHoff, 'N/A', fittingParameters['error']/(4.184*4.184), fitData['rmsr']/4.184, vantHoff/meltingTemp, '1.0'] elif model == 'NonTwoState': #Parameters A, Tm and Calorimetric startValues = [1, meltingGuess, enthalpyGuess] fittingParameters,fittingInstance=Fitting.doFit(expdata=zip(dscCurve.temperature,value), model='DSCindependent', guess=False, startvalues=startValues, silent=True) fitData = fittingInstance.getResult() vantHoff = fitData['deltaH']/4.184 calorimetric = fitData['A']*vantHoff meltingTemp = fitData['Tm'] res = ['NonTwoState', method, self.foldedRange[1], self.unfoldedRange[0], meltingTemp, vantHoff, calorimetric, fittingParameters['error']/(4.184*4.184), fitData['rmsr']/4.184, vantHoff/meltingTemp, calorimetric/vantHoff] elif model == 'Irreversible': #Parameters Tm, Calorimetirc and Ea startValues = [meltingGuess, enthalpyGuess, 50] fittingParameters,fittingInstance=Fitting.doFit(expdata=zip(dscCurve.temperature,value), model='DSC2stateIrreversible', guess=False, startvalues=startValues, silent=True) fitData = fittingInstance.getResult() activationEnergy = fitData['E']/4.184 calorimetric = fitData['deltaH']/4.184 meltingTemp = fitData['Tm'] res = ['Irreversible', method, self.foldedRange[1], self.unfoldedRange[0], meltingTemp, calorimetric, activationEnergy, fitData['error']/(4.184*4.184), fitData['rmsr']/4.184, calorimetric/meltingTemp] else: raise ValueError, 'Unknown DSC model %s' % model return [res, fittingParameters['error']/(4.184*4.184)]
def fit(self, model):
'''Fits the normalised curve to the specified dsc model
Params:
model - A string. Valid values are TwoState, NonTwoState and Irreversible.
Invalid values cause a ValueError exception to be raised
Returns:
A tuple. The first element is a list containing the fit data.
Strings decribing each element can be obtained by calling fitHeaders(model).
The second element is the fit error'''
if len(self.unfoldedRegion['temperatures']
) < 2 or self.useFolded is True:
method = 'FoldedOnly'
else:
method = 'Both'
dscCurve = self.normalisedCurve()
#conver to kj - fitting module has R in kj
value = [el * 4.184 for el in dscCurve.value]
#Use area under curve (calculated in progressBaseline) as enthalpy guess
#Melting temp guess is the temperature with max cp
enthalpyGuess = self.curveArea * 4.184
meltingGuess = dscCurve.temperature[value.index(max(value))]
if model == 'TwoState':
#Parameters Tm, VantHoff
startValues = [meltingGuess, enthalpyGuess]
fittingParameters, fittingInstance = Fitting.doFit(
expdata=zip(dscCurve.temperature, value),
model='DSC2state',
guess=False,
startvalues=startValues,
silent=True)
#Use fitData since it gives dict entries names - increased readibility
fitData = fittingInstance.getResult()
vantHoff = fitData['deltaH'] / 4.184
meltingTemp = fitData['Tm']
res = [
'TwoState', method, self.foldedRange[1], self.unfoldedRange[0],
meltingTemp, vantHoff, 'N/A',
fittingParameters['error'] / (4.184 * 4.184),
fitData['rmsr'] / 4.184, vantHoff / meltingTemp, '1.0'
]
elif model == 'NonTwoState':
#Parameters A, Tm and Calorimetric
startValues = [1, meltingGuess, enthalpyGuess]
fittingParameters, fittingInstance = Fitting.doFit(
expdata=zip(dscCurve.temperature, value),
model='DSCindependent',
guess=False,
startvalues=startValues,
silent=True)
fitData = fittingInstance.getResult()
vantHoff = fitData['deltaH'] / 4.184
calorimetric = fitData['A'] * vantHoff
meltingTemp = fitData['Tm']
res = [
'NonTwoState', method, self.foldedRange[1],
self.unfoldedRange[0], meltingTemp, vantHoff, calorimetric,
fittingParameters['error'] / (4.184 * 4.184),
fitData['rmsr'] / 4.184, vantHoff / meltingTemp,
calorimetric / vantHoff
]
elif model == 'Irreversible':
#Parameters Tm, Calorimetirc and Ea
startValues = [meltingGuess, enthalpyGuess, 50]
fittingParameters, fittingInstance = Fitting.doFit(
expdata=zip(dscCurve.temperature, value),
model='DSC2stateIrreversible',
guess=False,
startvalues=startValues,
silent=True)
fitData = fittingInstance.getResult()
activationEnergy = fitData['E'] / 4.184
calorimetric = fitData['deltaH'] / 4.184
meltingTemp = fitData['Tm']
res = [
'Irreversible', method, self.foldedRange[1],
self.unfoldedRange[0], meltingTemp, calorimetric,
activationEnergy, fitData['error'] / (4.184 * 4.184),
fitData['rmsr'] / 4.184, calorimetric / meltingTemp
]
else:
raise ValueError, 'Unknown DSC model %s' % model
return [res, fittingParameters['error'] / (4.184 * 4.184)]
