• Adds checks to confirm the correct COBRAme version is being used

• Improve Docker builds

• Correct bug causing 3 membrane complex stoichiometries to be overridden.

ECOLIme

Add compartments to iJL1678b_ME

• Added ecolime/util/helper_functions.py and ecolime/compartments.py to add compartments to all metabolites
• Fixes #19

Model corrections

• ADCLY-CPLX should be modified by pydx5p not pydx

• CYT-O-UBIOX-CPLX should have hemeO modification as well

• Masses for biomass constraints are computed from formula using metabolite.formula_weight this led to very minor changes in biomass stoichoimetries

• Glycyl was mispelled in keff.json causing these keffs not to be set
  correctly

  • Fixes #18

COBRAme

Adds support for Docker

• Dockerfiles added which can install COBRAme, ECOLIme, solvemepy, soplex_cython and a compiled version of qMINOS and SoPlex. It also installs all of the dependencies and builds iJL1678b-ME

• A Docker image containing all of the above can now be obtained from DockerHub .

Adds compartments to ME-model JSONs and refactored JSON saving/loading scripts

• Save/load full ME-model with:

from cobrame.io.json import save_json_me_model, load_json_me_model
save_json_me_model(me_model, '[save_loc]/model.json')
new_me_model = load_json_me_model('[save_loc]/model.json')

• Save/load reduced versino of ME-model with:

from cobrame.io.json import save_reduced_json_me_model, load_reduced_json_me_model
save_reduced_json_me_model(me_model, '[save_loc]/model.json')
new_me_model = load_reduced_json_me_model('[save_loc]/model.json')

Mass is no longer a property of model metabolites

• Now uses formula_weight and calculates mass based on formula

Additional changes

• Updated documentation to include examples of coupling constraint derivations

• Corrected mod_glycl id to mod_glycyl
• Added iron sulfur cluster to QueG complex
• Added full JSON dumping to end of build scripts

To avoid confusion with future E. coli ME-model releases, the output of the model building process included in ECOLIme will be renamed from iLE1678-ME to iJL1678b-ME. There is no new genetic content added to the model. The major differences between iJL1678-ME and iJL1678b-ME are:

1. The model structure has been reformulated to the COBRAme structure
2. Membrane surface area constraints are not included in the model
3. The model has been corrected and gap filled when necessary

The specific changes included in this release are described below:

ECOLIme

ECOLIme (iJL1678b) Changes

1. Reformulated biomass constraint

   • Add biomass constraint types in one model method as well as conversion of biomass types to global "biomass"
   • This requires a "biomass_type" attribute for PosttranslationData to know what biomass constraint to create as a result of modifications from SubreactionData. The JSONSCHEMA and json loading/dumping was modified to reflect this change.

2. Parameter changes

   • m_rr listed on bionumbers was a significant overestimation for e. coli, therefore it was changed to 1453 based on summing the average 5s, 16s and 23s rRNA molecular weights which were 38.4, 491.1, and 923.9,
     respectively
   • o2 exchange, gam, gam and unmodeled protein was changed to iJL1678 default parameters as the model transitioned to iJL1678b

3. Misc Corrections

   • Biomass contribution of the b1677 lipoprotein was removed from core_structural_demand_brauns reaction since this is accounted for in translation and lipid modification reactions.

   • Only lipids from the iJO1366 core biomass composition should have demand reactions. This was corrected.

   • Updated corrections.py to prevent rpoH from acting as a sigma factor for transcribing stable RNAs.

   • Fix bug causing biomass to be incorrectly calculated for 7 complex formation reactions (reactions listed in 3e1b643).

   • Fixed bug creating RNA_degradosome as metabolite and not complex

   • Removed all mod_palmitate reactions since this is replaced by lipoprotein

   • Transcription Reactions

     • water not required for nucleotide polymerization

   • Translation Reactions

     • Water not needed for selenocysteine modification
     • PrfC chain release factor is a GTPase, this has been updated in the model
     • ATP cost of charging was underrepresented. To correct this, ATP->AMP cost is included as part of the charged_tRNA addition subreactions.
     • Fix bug causing ATP hydrolyzing components to be updated incorrectly for RNA Degradatation
     • Methionine Cleavage releases a hydrogen, this has been updated

   • Ribosome Synthesis

     • Era_dim (assembly factor) is a GTPase, gdp has been added to its stoichiometry

   • tRNA Charging

     • Mass balance checked tRNA modifications, resulting in some stoichiometry changes
     • CHxxx_c metabolites no longer required
     • Updated electron acceptors with THIOREDOXINs etc.
     • The two lysine tRNAs can use two different synthetases, the synthetase to amino acid mapping has been updated with the generic form (generic_LYSINEaaRS) as in iOL1650
     • Correct mnm5U element_contribution and update cmo5U element_contribution and stoichiometry. Cmo5U begins with a conversion to 5-hydroxyuridine, which was unaccounted for an has an unknown mechanism. Water/protons were added to the stoichiometry to mass balance the addition of the hydroxyl group.

   • Translocation Reactions

     • Corrected by adding Water and hydrogen to ATP hydrolysis

   • Metabolic Reactions

     • Added reactions to synthesize cu_c (CU2R) and NiFeCoCN2_c (mod_NiFeCoCN2_c).
     • Removed reactions to uptake these metabolites at no metabolic cost
     • Removed CITLY-CPLX reaction since it had no effect on model
     • Replaced PFL_act reaction with mod_glycl modification for all PFL isozymes.
     • Removed hydrogens from iron-sulfur forming reactions since the net effect of zero hydrogen production or consumption

   • ComplexFormation Reactions

     • mod_coo is not a modification, it is the normal unmodified protein. All instances of EG11597-MONOMER_mod_coo have been replaced with EG11597-MONOMER
     • Corrected ATPS4r isozyme complex stoichiometry
     • Adding formulas increases the biomass produced in complex formation reactions for modification metabolites that previously did not have a formula (ie. pqq_c)
     • H+ moved to products of mod_lipoalt stoichiometry

4. ECOLIme Refactoring

   • Removed ecoli_k12.py and moved information into relevant files.
   • Deleted amino_acid_tRNA_synthetase.json moved it to tRNA_charging.py
   • Moved global_info into *.py file if possible. Goal to reduce the global_info fields as much as possible.
   • Removed rho_dependant from transcription_data attributes. This is bacteria specific.
   • Cleaned up DNA_replication.py and reaction

New ECOLIme Features

1. Mass Balance Testing (1715e36)

   • Added functions for assigning formulas to macromolecules (RNA, protein, complexes, etc.)
   • Added test for mass balance checking
   • Each SubreactionData transformation must be mass balanced by confirming that the net chemical gain/loss as a result of the subreaction is equivalent to user defined SubreactionData.set_elemental_contribution

2. MIsc. Features

   • Allow options in corrections.py to enable production of iOL1650 equivalent model.
   • Prepare for compatibility with new COBRApy versions
   • Model corrections (from iOL1650) compiled and and included in newly added corrections.py

COBRAme

Cobrame changes

1. COBRAme Refactoring

   • PEP8 compliance, wild card imports removed, modules renamed to be PEP8 compliant
   • Removed complex_data.chaperones. Chaperone activity should be handled as a subreaction and currently had no functionality.
   • _update_parent_reactions was changed to a public function

2. Model Structure Changes

   • Previously, a MEModel object contains both a process_data dictlist along with dictlists for each individual ProcessData type. This is redundant and can lead to confusing behavior. The individual ProcessData dictlists were removed and replaced with getters that return a generator containing all instances of the specific ProcessData type
   • PostTranslationReaction was updated to include biomass contributions of subreactions (ie. lipid modifications for lipoproteins)
   • The translocation pathway usage has no value associated with them. Therefore this was refactored to include a set of the pathways used, not a dict. The JSONSCHEMA and json dumping/loading was edited to reflect this.
   • Functionally, ModificationData and SubreactionData are identical. This lead to a lot of redundant and inefficient code. COBRAme was refactored to remove ModificationData as a ProcessData class. Moving forward modifications will be represented as SubreactionData and will be denoted by the “mod_” prefix in the ID.
   • Move subreaction attributes + translocation_multipliers from global info. Global info should be reserved for true global parameteris (i.e. m_rr, f_tRNA, etc.)
   • Removed rho_dependant from transcription_data attributes. This is bacteria specific.

3. COBRAme Bug Fixes

   • Corrected typo from coupling constraint formulation for k_mrna

New COBRAme Features

1. Added JSONSCHEMA and loading/dumping of fully functional ME-models

   • Using cobrame.io.json.save_full_me_model()

2. Mass Balance Handling

   • Added formulas.py
     • Functions to add formulas for all complexes produced from metabolic reactions based mostly on formulas in modifications.txt

3. Improved Documentation

   • Most of core module completely documented
   • Updated tutorial and readthedocs

4. Misc new features

   • Add SubreactionData.get_all_usages() method to yield all instances where a process_data uses the subreaction.
   • Prepare for compatibility with new COBRApy versions
   • Add setters for elemental_contribution in SubreactionData and stoichiometry in StoichiometricData

• Moved biomass allocation characterization functions to ecolime
• added get_SASA_keffs function to set SASA based keff approximations
• bug fixes for determining translocate costs
• moved translation and transcription building functions to dedicated scripts
• added function for calculating gr dependent DNA concentrations as in iOL1650
• functions for MPI parallel simulation

1)Reformulated surface area capacity constraint to ensure unit consistency
2)Added protein/nonprotein segregation to inner/outer membrane
3)Updated parameter sweeps to enable probing of linear acetate overflow metabolism trends
4)Move translocation formation reactions for prototype notebook to translocation.py
5)Correction of keff temperature scaling and small protein groEL folding
6)Replaced dummy_protein with dummy_protein_biomass for unmodeled protein dilution reaction
7)Corrected bug that allowed model to cheat by overproducing mRNA biomass. This should allow the enyzme efficiency increase with growth rate effect
8)Updated binary search to allow minimization of enzyme_efficiency_scaling
9)Models can now be saved as a json