Performing FVA in MICOM

[sankarmaheshr]

Hi. Is there a way to calculate the maximum possible flux through a reaction (FVA) in MICOM when building a community model?

[justuswfink]

Hi there! As far as I know, it is not directly possible in the package itself but you can do it manually. Have you seen question #138 in this discussion? The answer by Christian there shows a concise way how to it.

In a more lengthy way, the following has worked for me. Basically, the idea is to get the growth rates from MICOM, add them as constraints to the model and then pass the same object to the built-in FVA analysis in CobraPy.

import numpy as np
from cobra.flux_analysis import flux_variability_analysis
from micom import Community, data

# let's build a 2-strain community using metabolic models that are provided with the package
taxa = data.test_taxonomy()
subset = [0,1] # pick strains
com = Community(taxa.loc[subset])

#then optimize as usual, using MICOM. 
result = com.cooperative_tradeoff()
growth_rates = result.members.growth_rate # need to store growth rates for later

### add constraints for each strain to MICOM object
rate = growth_rates.loc['Escherichia_coli_1']                          # this is the id of our strain at taxa.iloc[0]
rxn  = 'BIOMASS_Ecoli_core_w_GAM__Escherichia_coli_1'
com.reactions.get_by_id(rxn).lower_bound = rate
com.reactions.get_by_id(rxn).upper_bound = rate

rate = growth_rates.loc['Escherichia_coli_2']                        # this is the id of our strain at taxa.iloc[1]
rxn  = 'BIOMASS_Ecoli_core_w_GAM__Escherichia_coli_2'
com.reactions.get_by_id(rxn).lower_bound = rate
com.reactions.get_by_id(rxn).upper_bound = rate

# we then call the flux variability analysis from COBRA on the MICOM object
## let's focus on the exchange reaction between species and for the medium
fluxes = flux_variability_analysis(com, com.internal_exchanges + com.exchanges)

## as a sanity check, we can also ask for flux variability of our biomass reactions
# we expect zero variation here, since we have explicitly constrained them before
flux_variability_analysis(com, ['BIOMASS_Ecoli_core_w_GAM__Escherichia_coli_1', 'BIOMASS_Ecoli_core_w_GAM__Escherichia_coli_2'])

Here is an example of the output I get for this. Fluxes with underscore _m show the absorption of resources from the external source into the medium, and fluxes with underscore _e show the absorption of resources by the strains from the medium. Negative values here mean intake of resources, this is a standard convention of FBA as I learned.

The catch with FVA on larger community models is that the runtime, as pointed out by Christian earlier (see again question #138). The FVA here for these small metabolic models runs 20s for all 60 reactions, but this could quickly get out of hand with a community, so you will have to try and see if it is still feasible.

Justus

[cdiener]

Thanks @justuswfink, perfect explanation. For setting the growth rate constraints from a previous solution that can also be done a bit faster with _apply_min_growth.

In [1]: import micom as mm

In [2]: from cobra.flux_analysis import flux_variability_analysis

In [3]: tax = mm.data.test_taxonomy()

In [4]: com = mm.Community(tax)
Building ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 100% 0:00:00

In [5]: sol = com.cooperative_tradeoff(fraction=1.0)

In [6]: with com:
   ...:     mm.util._apply_min_growth(com, sol.members.growth_rate.drop("medium"))
   ...:     fva = flux_variability_analysis(com, com.exchanges)
   ...: 

In [7]: fva
Out[7]: 
                  minimum       maximum
EX_ac_m      4.684868e-16  5.880137e-14
EX_acald_m   4.726860e-13  4.181389e-15
EX_akg_m     5.415208e-14  2.427597e-10
EX_co2_m     2.280983e+01  2.280983e+01
EX_etoh_m    3.914466e-14  1.154832e-14
EX_for_m     2.142067e-15  2.195611e-09
EX_fru_m     0.000000e+00  0.000000e+00
EX_fum_m     0.000000e+00  0.000000e+00
EX_glc__D_m -1.000000e+01 -1.000000e+01
EX_gln__L_m  0.000000e+00  0.000000e+00
EX_glu__L_m  8.795124e-14  4.649916e-11
EX_h_m       1.753087e+01  1.753087e+01
EX_h2o_m     2.917583e+01  2.917583e+01
EX_lac__D_m  2.367835e-14  1.367713e-10
EX_mal__L_m  0.000000e+00  0.000000e+00
EX_nh4_m    -4.765319e+00 -4.765319e+00
EX_o2_m     -2.179949e+01 -2.179949e+01
EX_pi_m     -3.214895e+00 -3.214895e+00
EX_pyr_m     3.901137e-14  2.478758e-10
EX_succ_m    1.156386e-15  7.443959e-15