def parse_fidarestat_data():
"""
Convert fidarestat data into simple text file, with columns
gene = Ensembl
C = non-diabetic controls
D = diabetic untreated
F = diabetic + fidarestat
"""
book = xlrd.open_workbook(os.path.join(PATH, 'fidarestat.xls'))
sbml = om.read_sbml(os.path.join(PATH, RAMBO_NAME))
gene_list = om.get_list_of_genes(sbml)
sheet = book.sheet_by_name("maxd1antilog2")
file = os.path.join(PATH, 'fidarestat.txt')
with open(file, 'wb') as csvfile:
writer = csv.writer(csvfile, delimiter=',')
writer.writerow(['gene', 'C', 'D', 'F'])
done = []
for row in range(1, sheet.nrows):
F = strip_string(sheet.cell(row, 3).value)
C = strip_string(sheet.cell(row, 4).value)
D = strip_string(sheet.cell(row, 5).value)
gene = strip_string(sheet.cell(row, 17).value)
if (gene in gene_list) and (gene not in done):
done.append(gene)
writer.writerow([gene, C, D, F])
for gene in gene_list:
if gene not in done:
done.append(gene)
writer.writerow([gene, 0, 0, 0])
def parse_timecourse_data():
"""
Convert timecourse data into simple text file, with columns
gene = Ensembl
C4 = controls after 4 weeks
D4 = diabetic after 4 weeks
C8 = controls after 4 weeks
D8 = diabetic after 4 weeks
"""
book = xlrd.open_workbook(os.path.join(PATH, 'timecourse.xls'))
sbml = om.read_sbml(os.path.join(PATH, RAMBO_NAME))
gene_list = om.get_list_of_genes(sbml)
sheet = book.sheet_by_name("E-MEXP-515-processed-data-13435")
file = os.path.join(PATH, 'timecourse.txt')
all = np.array([])
with open(file, 'wb') as csvfile:
writer = csv.writer(csvfile, delimiter=',')
writer.writerow(['gene', 'C4', 'D4', 'C8', 'D8'])
done = []
data = []
for row in range(2, sheet.nrows):
C4 = strip_string(sheet.cell(row, 9).value)
if not C4:
C4 = 0
D4 = strip_string(sheet.cell(row, 17).value)
if not D4:
D4 = 0
C8 = strip_string(sheet.cell(row, 25).value)
if not C8:
C8 = 0
D8 = strip_string(sheet.cell(row, 33).value)
if not D8:
D8 = 0
gene = strip_string(sheet.cell(row, 1).value)
if (gene in gene_list) and (gene not in done):
done.append(gene)
writer.writerow([gene, C4, D4, C8, D8])
for X in [C4, D4, C8, D8]:
x = float(X)
if x:
data.append(x)
for gene in gene_list:
if gene not in done:
done.append(gene)
writer.writerow([gene, 0, 0, 0, 0])
print 'data range:\nmin:\t%g\nmean:\t%g\nmax:\t%g\n' %(np.min(data), np.mean(data), np.max(data))
def parse_liver_data():
"""
Convert liver data into simple text file, with columns
gene = Ensembl
C = non-diabetic controls
D = diabetic untreated
CT = control + drug
DT = diabetic + drug
Csd = control standard deviation
.. etc
"""
book = xlrd.open_workbook(os.path.join(PATH, 'liver.xls'))
sbml = om.read_sbml(os.path.join(PATH, RAMBO_NAME))
gene_list = om.get_list_of_genes(sbml)
sheet1 = book.sheet_by_name("Means")
sheet2 = book.sheet_by_name("Stdevs")
file = os.path.join(PATH, 'liver.txt')
with open(file, 'wb') as csvfile:
writer = csv.writer(csvfile, delimiter=',')
writer.writerow(['gene', 'C', 'Csd', 'D', 'Dsd', 'CT', 'CTsd', 'DT', 'DTsd'])
done = []
for row in range(1, sheet1.nrows):
gene = strip_string(sheet1.cell(row, 0).value)
D = strip_string(sheet1.cell(row, 1).value)
C = strip_string(sheet1.cell(row, 2).value)
DT = strip_string(sheet1.cell(row, 3).value)
CT = strip_string(sheet1.cell(row, 4).value)
Dsd = strip_string(sheet2.cell(row, 1).value)
Csd = strip_string(sheet2.cell(row, 2).value)
DTsd = strip_string(sheet2.cell(row, 3).value)
CTsd = strip_string(sheet2.cell(row, 4).value)
if (gene in gene_list) and (gene not in done):
done.append(gene)
writer.writerow([gene, C, Csd, D, Dsd, CT, CTsd, DT, DTsd])
for gene in gene_list:
if gene not in done:
done.append(gene)
writer.writerow([gene, 0, 0, 0, 0, 0, 0, 0, 0])
def liver_to_flux():
"""
Convert liver data into flux using SuperDaaaaave
"""
sbml = om.read_sbml(os.path.join(PATH, RAMBO_NAME))
# set minimal glucose medium
media = [
'EX_ca2(e)', 'EX_cl(e)', 'EX_fe2(e)', 'EX_fe3(e)', 'EX_h(e)', 'EX_h2o(e)',
'EX_k(e)', 'EX_na1(e)', 'EX_nh4(e)', 'EX_so4(e)', 'EX_pi(e)', 'EX_o2(e)'
]
block_all_imports(sbml)
# set_import_bounds(sbml, 'EX_glc(e)', 1)
sbml.getModel().getReaction('R_EX_glc_LPAREN_e_RPAREN_').getKineticLaw().getParameter('LOWER_BOUND').setValue(1.)
sbml.getModel().getReaction('R_EX_glc_LPAREN_e_RPAREN_').getKineticLaw().getParameter('UPPER_BOUND').setValue(1.)
set_import_bounds(sbml, media, INF)
change_objective(sbml, 'DM_atp_c_')
# v, atp_max = om.optimize_cobra_model(sbml)
atp_max = 31.
# force at least 50% of max atp generation
sbml.getModel().getReaction('R_DM_atp_c_').getKineticLaw().getParameter('LOWER_BOUND').setValue(atp_max/2.)
# limit sorbitol dehydrogenase
sbml.getModel().getReaction('R_SBTD_D2').getKineticLaw().getParameter('UPPER_BOUND').setValue(1.)
# allow sorbitol export to sink
sbml.getModel().getReaction('R_SBTle').getKineticLaw().getParameter('LOWER_BOUND').setValue(-INF)
# create gene data
gene_names, gene_exp_C, gene_exp_D, gene_exp_CT, gene_exp_DT, = [], [], [], [], []
gene_exp_sd_C, gene_exp_sd_D, gene_exp_sd_CT, gene_exp_sd_DT, = [], [], [], []
file = os.path.join(PATH, 'liver.txt')
with open(file, 'rU') as csvfile:
reader = csv.DictReader(csvfile, delimiter=',')
for row in reader:
gene = row['gene']
C = row['C']
D = row['D']
CT = row['CT']
DT = row['DT']
Csd = row['Csd']
Dsd = row['Dsd']
CTsd = row['CTsd']
DTsd = row['DTsd']
gene_names.append(gene)
gene_exp_C.append(float(C))
gene_exp_D.append(float(D))
gene_exp_CT.append(float(CT))
gene_exp_DT.append(float(DT))
gene_exp_sd_C.append(float(Csd))
gene_exp_sd_D.append(float(Dsd))
gene_exp_sd_CT.append(float(CTsd))
gene_exp_sd_DT.append(float(DTsd))
gene_exp_C, gene_exp_D, gene_exp_CT, gene_exp_DT = np.array(gene_exp_C), np.array(gene_exp_D), np.array(gene_exp_CT), np.array(gene_exp_DT)
gene_exp_sd_C, gene_exp_sd_D, gene_exp_sd_CT, gene_exp_sd_DT = np.array(gene_exp_sd_C), np.array(gene_exp_sd_D), np.array(gene_exp_sd_CT), np.array(gene_exp_sd_DT)
# but set zero means as small sd
for gene_exp_sd_X in [gene_exp_sd_C, gene_exp_sd_D, gene_exp_sd_CT, gene_exp_sd_DT]:
gene_exp_sd_X[gene_exp_sd_X == 0] = min(gene_exp_sd_X[gene_exp_sd_X != 0])/2
USE_SD = False # ignore SDs: assume all data has equal weighting
flux_C, scaling_factor_C = SuperDaaaaave(sbml, gene_names, gene_exp_C, gene_exp_sd_C, MaxGrowth=False, UseSD=USE_SD, FixScaling=0)
flux_CT, scaling_factor_CT = SuperDaaaaave(sbml, gene_names, gene_exp_CT, gene_exp_sd_CT, MaxGrowth=False, UseSD=USE_SD, FixScaling=scaling_factor_C, TargetFlux=flux_C)
# fix glucose uptake in diabetic & treated at 6x level in control
# set_import_bounds(sbml, 'EX_glc(e)', 6.)
sbml.getModel().getReaction('R_EX_glc_LPAREN_e_RPAREN_').getKineticLaw().getParameter('LOWER_BOUND').setValue(10.)
sbml.getModel().getReaction('R_EX_glc_LPAREN_e_RPAREN_').getKineticLaw().getParameter('UPPER_BOUND').setValue(10.)
flux_D, scaling_factor_D = SuperDaaaaave(
sbml, gene_names, gene_exp_D, gene_exp_sd_D, MaxGrowth=False,
UseSD=USE_SD, FixScaling=scaling_factor_C, TargetFlux=flux_C
)
flux_DT, scaling_factor_DT = SuperDaaaaave(
sbml, gene_names, gene_exp_DT, gene_exp_sd_DT, MaxGrowth=False,
UseSD=USE_SD, FixScaling=scaling_factor_C, TargetFlux=flux_D
)
# remove small fluxes
flux_C, flux_D, flux_CT, flux_DT = np.array(flux_C), np.array(flux_D), np.array(flux_CT), np.array(flux_DT)
for X in [flux_C, flux_D, flux_CT, flux_DT]:
X[abs(X) < 1e-6] = 0
# write results
file = os.path.join(PATH, 'om_liver.txt')
with open(file, 'wb') as csvfile:
writer = csv.writer(csvfile, delimiter=',')
writer.writerow(['rxn_id', 'C', 'D', 'CT', 'DT'])
for index, reaction in enumerate(sbml.getModel().getListOfReactions()):
rID = reaction.getId()
C = flux_C[index]
D = flux_D[index]
CT = flux_CT[index]
DT = flux_DT[index]
writer.writerow([rID, C, D, CT, DT])
def timecourse_to_flux():
"""
Convert timecourse data into flux using SuperDaaaaave
"""
sbml = om.read_sbml(os.path.join(PATH, RAMBO_NAME))
# set minimal glucose medium
media = [
'EX_ca2(e)', 'EX_cl(e)', 'EX_fe2(e)', 'EX_fe3(e)', 'EX_h(e)', 'EX_h2o(e)',
'EX_k(e)', 'EX_na1(e)', 'EX_nh4(e)', 'EX_so4(e)', 'EX_pi(e)', 'EX_o2(e)'
]
block_all_imports(sbml)
# set_import_bounds(sbml, 'EX_glc(e)', 1)
sbml.getModel().getReaction('R_EX_glc_LPAREN_e_RPAREN_').getKineticLaw().getParameter('LOWER_BOUND').setValue(1.)
sbml.getModel().getReaction('R_EX_glc_LPAREN_e_RPAREN_').getKineticLaw().getParameter('UPPER_BOUND').setValue(1.)
set_import_bounds(sbml, media, INF)
change_objective(sbml, 'DM_atp_c_')
# v, atp_max = om.optimize_cobra_model(sbml)
atp_max = 31.
# force at least 50% of max atp generation
sbml.getModel().getReaction('R_DM_atp_c_').getKineticLaw().getParameter('LOWER_BOUND').setValue(atp_max/2.)
# limit sorbitol dehydrogenase
sbml.getModel().getReaction('R_SBTD_D2').getKineticLaw().getParameter('UPPER_BOUND').setValue(1.)
# allow sorbitol export to sink
sbml.getModel().getReaction('R_SBTle').getKineticLaw().getParameter('LOWER_BOUND').setValue(-INF)
# create gene data
gene_names, gene_exp_C4, gene_exp_D4, gene_exp_C8, gene_exp_D8, = [], [], [], [], []
file = os.path.join(PATH, 'timecourse.txt')
with open(file, 'rU') as csvfile:
reader = csv.DictReader(csvfile, delimiter=',')
for row in reader:
gene = row['gene']
C4 = row['C4']
D4 = row['D4']
C8 = row['C8']
D8 = row['D8']
gene_names.append(gene)
gene_exp_C4.append(float(C4))
gene_exp_D4.append(float(D4))
gene_exp_C8.append(float(C8))
gene_exp_D8.append(float(D8))
gene_exp_C4, gene_exp_D4 = np.array(gene_exp_C4), np.array(gene_exp_D4)
gene_exp_C8, gene_exp_D8 = np.array(gene_exp_C8), np.array(gene_exp_D8)
# no sds given, assume equal to mean
gene_exp_sd_C4, gene_exp_sd_D4 = gene_exp_C4.copy(), gene_exp_D4.copy()
gene_exp_sd_C8, gene_exp_sd_D8 = gene_exp_C8.copy(), gene_exp_D8.copy()
# but set zero means as small sd
for gene_exp_sd_X in [gene_exp_sd_C4, gene_exp_sd_D4, gene_exp_sd_C8, gene_exp_sd_D8]:
gene_exp_sd_X[gene_exp_sd_X == 0] = min(gene_exp_sd_X[gene_exp_sd_X != 0])/2
USE_SD = False # ignore SDs: assume all data has equal weighting
flux_C4, scaling_factor_C4 = SuperDaaaaave(sbml, gene_names, gene_exp_C4, gene_exp_sd_C4, MaxGrowth=False, UseSD=USE_SD, FixScaling=0)
flux_C8, scaling_factor_C8 = SuperDaaaaave(sbml, gene_names, gene_exp_C8, gene_exp_sd_C8, MaxGrowth=False, UseSD=USE_SD, FixScaling=scaling_factor_C4, TargetFlux=flux_C4)
# fix glucose uptake in diabetic & treated at 10x level in control
sbml.getModel().getReaction('R_EX_glc_LPAREN_e_RPAREN_').getKineticLaw().getParameter('LOWER_BOUND').setValue(10.)
sbml.getModel().getReaction('R_EX_glc_LPAREN_e_RPAREN_').getKineticLaw().getParameter('UPPER_BOUND').setValue(10.)
flux_D4, scaling_factor_D4 = SuperDaaaaave(
sbml, gene_names, gene_exp_D4, gene_exp_sd_D4, MaxGrowth=False,
UseSD=USE_SD, FixScaling=scaling_factor_C4, TargetFlux=flux_C4
)
flux_D8, scaling_factor_D8 = SuperDaaaaave(
sbml, gene_names, gene_exp_D8, gene_exp_sd_D8, MaxGrowth=False,
UseSD=USE_SD, FixScaling=scaling_factor_C8, TargetFlux=flux_C8
)
# remove small fluxes
flux_C4, flux_D4 = np.array(flux_C4), np.array(flux_D4)
flux_C8, flux_D8 = np.array(flux_C8), np.array(flux_D8)
for X in [flux_C4, flux_D4, flux_C8, flux_D8]:
X[abs(X) < 1e-6] = 0
# write results
file = os.path.join(PATH, 'om_timecourse.txt')
with open(file, 'wb') as csvfile:
writer = csv.writer(csvfile, delimiter=',')
writer.writerow(['rxn_id', 'C4', 'D4', 'C8', 'D8'])
for index, reaction in enumerate(sbml.getModel().getListOfReactions()):
rID = reaction.getId()
C4 = flux_C4[index]
D4 = flux_D4[index]
C8 = flux_C8[index]
D8 = flux_D8[index]
writer.writerow([rID, C4, D4, C8, D8])
