def row2string(S_row, cids):
active_cids = list(np.nonzero(S_row)[0].flat)
sparse = dict((cids[c], S_row[c])
for c in active_cids
if abs(S_row[c]) > 1e-10)
r = Reaction("", sparse)
return r.FullReactionString(show_cids=False)
['C%05d' % all_cids[c],
'%8.1f' % dG0_f[c, 0], 'Calculated'])
#figure()
#plot(b, dot(S_red, dG0_f[unknown_columns]), '.')
figure()
plot(dG0_r[known_rows], dot(S[known_rows, :], dG0_f), '.')
show()
csv_out = csv.writer(open('../res/acetogens_reactions.csv', 'w'))
csv_out.writerow([
"RID", "EC", "REACTION", "dG0 measured", "dG0 calculated", "pH", "I", "T"
])
for r in xrange(len(reactions)):
(rid, ec, sparse, dG0, pH, I, T) = reactions[r]
reaction = Reaction(ec, sparse, rid=rid)
dG0_calculated = dot(S[r, :], dG0_f)
if r in known_rows:
csv_out.writerow([
rid, ec,
reaction.FullReactionString(),
'%.1f' % dG0,
'%.1f' % dG0_calculated, pH, I, T
])
else:
csv_out.writerow([
rid, ec,
reaction.FullReactionString(), 'N/A',
'%.1f' % dG0_calculated, pH, I, T
])
#H.I = 0.25
#H.T = 300;
#sparse_reaction = {13:-1, 1:-1, 9:2}
#sparse_reaction = {36:-1, 3981:1}
#sparse_reaction = {6:-1, 143:-1, 234:1, 5:1}
#sparse_reaction = {1:-1, 499:-1, 603:1, 86:1}
#sparse_reaction = {1:-1, 6:-1, 311:-1, 288:1, 5:1, 80:2, 26:1}
#sparse_reaction = {408:-1, 6:-1, 4092:1, 5:1}
#sparse_reaction = {588:-1, 1:-1, 114:1, 9:1}
#sparse_reaction = {1:-1, 3:-1, 149:-1, 288:1, 4:1, 80:2, 22:1}
react = Reaction("reaction", {408: -1, 6: -1, 4092: 1, 5: 1})
#sys.stdout.write("The dG0_r of PPi + H20 <=> 2 Pi: \n\n")
react.Balance()
print react.FullReactionString()
sys.stdout.write("%5s | %5s | %6s | %6s\n" % ("pH", "I", "T", "dG0_r"))
for pH in np.arange(5, 10.01, 0.25):
H_withpka.pH = pH
sys.stdout.write("%5.2f | %5.2f | %6.1f | %6.2f\n" %
(H_withpka.pH, H_withpka.I, H_withpka.T,
react.PredictReactionEnergy(H_withpka)))
for cid in react.get_cids():
print '-' * 50
print "C%05d - %s:" % (cid, H_withpka.kegg.cid2name(cid))
print H_withpka.kegg.cid2inchi(cid)
print "Pseudoisomers:\n", H_withpka.cid2PseudoisomerMap(cid)
print "dG0'_f = %.1f kJ/mol" % H_withpka.cid2PseudoisomerMap(
cid).Transform(pH=7, I=0, pMg=10, T=298.15)
def ConvertFormation2Reaction(self, output_fname):
logging.info("Converting all formation energies to reactions")
output_csv = csv.writer(open(output_fname, 'w'))
# keep the format used for TECRDB
output_csv.writerow(
('ref', 'ID', 'method', 'eval', 'EC', 'name', 'kegg_reaction',
'reaction', 'dG0\'', 'T', 'I', 'pH', 'pMg'))
atom2cid = {}
for atom, (name, stoich) in KeggObservation.ATOM2ELEMENT.iteritems():
cid, _, _ = self.kegg.name2cid(name, 0)
if cid is None:
raise Exception(
"Cannot find the element %s in the KEGG database" % name)
atom2cid[atom] = (cid, stoich)
#output_csv.writerow(('element',
# 'C%05d' % cid, 'formation', 'A', '',
# 'formation of %s' % self.kegg.cid2name(cid),
# "C%05d" % cid,
# name, 0, self.T, self.I, self.pH, self.pMg))
for label in ['training', 'testing']:
ptable = PsuedoisomerTableThermodynamics.FromCsvFile(
self.FormationEnergyFileName, label=label)
for cid in ptable.get_all_cids():
pmatrix = ptable.cid2PseudoisomerMap(cid).ToMatrix()
if len(pmatrix) != 1:
raise Exception("multiple training species for C%05d" %
cid)
nH, _charge, nMg, dG0 = pmatrix[0]
diss_table = dissociation.GetDissociationTable(cid, False)
if diss_table is None:
continue
diss_table.SetFormationEnergyByNumHydrogens(dG0, nH, nMg)
dG0_prime = diss_table.Transform(pH=self.pH,
I=self.I,
pMg=self.pMg,
T=self.T)
ref = ptable.cid2SourceString(cid)
atom_bag = self.kegg.cid2atom_bag(cid)
if not atom_bag:
continue
ne = self.kegg.cid2num_electrons(cid)
elem_ne = 0
sparse = {cid: 1}
for elem, count in atom_bag.iteritems():
if elem == 'H':
continue
elem_ne += count * Molecule.GetAtomicNum(elem)
elem_cid, elem_coeff = atom2cid[elem]
sparse.setdefault(elem_cid, 0)
sparse[elem_cid] += -count * elem_coeff
# use the H element to balance the electrons in the formation
# reactions (we don't need to balance protons since this is
# a biochemical reaction, so H+ are 'free').
H_cid, H_coeff = atom2cid['H']
sparse[H_cid] = (elem_ne - ne) * H_coeff
reaction = Reaction(
"formation of %s" % self.kegg.cid2name(cid), sparse)
output_csv.writerow(
(ref, 'C%05d' % cid, 'formation', 'A', '',
'formation of %s' % self.kegg.cid2name(cid),
reaction.FullReactionString(),
reaction.FullReactionString(show_cids=False),
'%.2f' % dG0_prime, self.T, self.I, self.pH, self.pMg))
