import sys
import string_chem_net as scn
import itertools as it
try:
(monos, max_pol, min_ins, max_ins, min_outs, max_outs, reps) = sys.argv[1:]
except ValueError:
sys.exit(
'Arguments:\nmonomers\nmax polymer length\nminimum ' +
'number of food sources\nmaximum number of food sources\nminimum ' +
'number of biomass precursors\nmaximum number of biomass precursors\n'
+ 'number of times to prune with each setup')
# create the universal network
SCN = scn.CreateNetwork(monos, int(max_pol))
universal_model = scn.make_cobra_model(SCN.met_list,
SCN.rxn_list,
allow_export=True)
# make every pair of number of environments and number of biomass precursors
conditions = list(
it.product(
range(int(min_ins),
int(max_ins) + 1), # have to add 1 because Python
range(int(min_outs),
int(max_outs) + 1)))
# store the reaction-to-metabolite ratio and the % of pruned reactions for each
# pruned network
output_data = list()
data.columns = ['env', 'rxn_incl', 'growth']
# add a column with the biomass components
data['biomass'] = list(
it.repeat('-'.join([met.id for met in bm_rxn.metabolites]),
len(food_mets)))
# reorder columns
data = data[['biomass', 'env', 'rxn_incl', 'growth']]
return (data)
monos = 'ab' # characters to use as monomers
max_len = 5 # maximum length of each string chemical
ins = 2 # number of food sources / environmental nutrients
envs = 50 # number of different sets of food sources per biomass reaction
outs = 5 # number of biomass precursors
orgs = 10 # number of different biomass reactions
combos = 50 # number of times to perturb coefficients per biomass reaction
threads = 4 # threads to use when pruning in parallel
SCN = scn.CreateNetwork(monos, max_len)
full_model = scn.make_cobra_model(SCN.met_list, SCN.rxn_list)
# prune network using many biomass reactions and environments
pool = mp.Pool(1)
data_bits = pool.map(
prune_many_times,
# same arguments every time for orgs times
[[full_model, ins, outs, envs, combos] for bm in range(orgs)])
data = pd.concat(data_bits)
data.to_csv('data/figure_S6_data.csv', index=False)
# return it
info_df = pd.DataFrame(info_dict)
return (info_df)
if __name__ == '__main__':
try:
(monos, max_len, ins, outs, reps) = sys.argv[1:]
except ValueError:
sys.exit(
'Specify monomers to use, maximum length of polymer, number of ' +
'food sources, number of biomass precursors, and number of times to '
+ 'prune')
# setup network of specified size
SCN = scn.CreateNetwork(monos, int(max_len))
cobra_model = scn.make_cobra_model(SCN.met_list, SCN.rxn_list)
# prepare dataframe to hold info from all rounds of pruning
all_data = pd.DataFrame({
'step': list(),
'type': list(),
'rxn_count': list(),
'jaccard': list()
})
# prune reps times and store all information in all_data
for i in range(int(reps)):
if (i + 1) % 10 == 0:
print(f'On rep {i+1} of {reps}')
some_data = compare_nets(cobra_model, int(ins), int(outs))
# make another column to put i in so we can separate out all the different
# figure_1,py
# makes the three example string chemistry graphs shown in figure 1 of the
# paper
import string_chem_net as scn
import pygraphviz as gv
import string
# tried making all three be subgraphs of the same graph, but the resulting
# image came out very wide and oddly low resolution, so we're making 3 separate
# images
# make the three networks we want to visualize, then loop over that list
least_net = scn.CreateNetwork('a', 2)
mid_net = scn.CreateNetwork('ab', 2)
most_net = scn.CreateNetwork('ab', 3)
nets = [least_net, mid_net, most_net]
for net in nets:
figure = gv.AGraph(
size = '11,8', # set size of output image
dpi = '600', # set resolution of output image
splines = 'true' # make sure nodes aren't overlapping with each other
)
# start with nodes
for met in net.met_list:
figure.add_node(met, shape = 'box', color = 'blue')
for rxn in net.rxn_list:
figure.add_node(rxn, shape = 'oval', color = 'red')
# then do edges
output = [ins, outs, ratio, pruned_pct]
# make everything a string so we can join it later
return ([str(x) for x in output])
# set parameters
monos = 'ab'
max_pol = 5
min_ins = 2 # smallest number of nutrients to use
max_ins = 5 # largest number of nutrients to use
min_outs = 5 # smallest number of biomass precursors to use
max_outs = 10 # largest number of biomass precursors to use
reps = 100 # number of times to prune with each combination
# create the universal network
SCN = scn.CreateNetwork(monos, max_pol)
universal_model = scn.make_cobra_model(SCN.met_list,
SCN.rxn_list,
allow_export=True)
# make every pair of number of environmental metabolites and number of biomass
# precursors
conditions = list(
it.product(
range(min_ins, max_ins + 1), # have to add 1 because Python
range(min_outs, max_outs + 1)))
# add the universal model to each of the sublists so that we can map this list
# of arguments to the function
args = [(universal_model, ) + sublist for sublist in conditions]
# multiply the list by reps to make sure each pair of values is tested reps
# times