Skip to main content
 首页 » 编程设计

python之如何将通量比设置为约束条件

2024年12月31日9mq0036

在某些数据集中,我有时会观察到固定的通量比,我想将其纳入我的模拟中。我如何在 CBMPy 中执行此操作?

例如,我有来自 here 的模型现在想将琥珀酸流出和丙酮酸流出的比率限制为 2.0。我知道如何对个人 react 设置限制:

import cbmpy 
 
# downloaded from http://bigg.ucsd.edu/models/e_coli_core 
ecoli = cbmpy.CBRead.readSBML3FBC('e_coli_core.xml') 
 
ecoli.setReactionBounds('R_EX_pyr_e', 1.0, 1000.0) 
ecoli.setReactionBounds('R_EX_succ_e', 2.0, 1000.0) 
 
# solve the model 
cbmpy.doFBA(ecoli) 
 
# get all reaction values 
solution = ecoli.getReactionValues() 
print(solution['R_EX_pyr_e']) 
print(solution['R_EX_succ_e']) 

对于这种情况,比率是正确的,但我如何将它添加为在所有条件下都满足的约束条件?

请您参考如下方法:

这确实是通量平衡分析 (FBA) 中的常用方法,您可以使用函数 addUserConstraint 来完成此操作。

整个代码示例可能如下所示(解释如下):

import cbmpy as cbm 
 
# downloaded from http://bigg.ucsd.edu/models/e_coli_core 
ecoli = cbm.CBRead.readSBML3FBC('e_coli_core.xml') 
 
# make a clone of the original model 
ecoli_ratio = ecoli.clone() 
 
# add the desired user constraint; explanation follows below 
ecoli_ratio.addUserConstraint("pyr_succ_ratio", fluxes=[(1.0, 'R_EX_pyr_e' ),(-0.5, 'R_EX_succ_e')], operator='=', rhs=0.0) 
 
# now we have to set only one flux bound; if you think it is naturally excreted, this step is not needed 
ecoli_ratio.setReactionBounds('R_EX_succ_e', 4.0, cbm.INF) 
 
cbm.doFBA(ecoli_ratio) 
solution = ecoli_ratio.getReactionValues() 
print("{}: {}".format("succinate excretion rate", solution['R_EX_succ_e'])) 
print("{}: {}".format("pyruvate excretion rate", solution['R_EX_pyr_e'])) 

这将打印

succinate excretion rate: 4.0 
pyruvate excretion rate: 2.0 

如您所见,比率为 2.0,符合要求。

更多解释:

约束是

J_succ / J_pyr = 2.0 

可以重写为

J_succ = 2.0 J_pyr 

最后

J_pyr - 0.5 J_succ = 0 

这正是我们在 addUserConstraint 中传递给 fluxes 的内容:

fluxes=[(1.0, 'R_EX_pyr_e' ),(-0.5, 'R_EX_succ_e')], operator='=', rhs=0.0) 

您可以通过打印来检查用户定义的约束:

print(ecoli_ratio.user_constraints) 
{'pyr_succ_ratio': {'operator': 'E', 'rhs': 0.0, 'fluxes': [(1.0, 'R_EX_pyr_e'), (-0.5, 'R_EX_succ_e')]}} 

因为这是一本字典,您可以通过简单地执行以下操作来删除约束:

del ecoli_ratio.user_constraints['pyr_succ_ratio'] 
print(ecoli_ratio.user_constraints) 
{} 

但我强烈建议您在每次对模型进行重大更改时创建一个克隆。