Here, we list the associated components in the example model, iPfal19, and discuss some of the challenges in implementing these standards. This is the third iteration of the Plasmodium falciparum 3D7 genome‐scale metabolic network reconstruction. The original reconstruction was generated using a custom pipeline and multiple rounds of curation were conducted (Carey et al, 2017). iPfal19 fails to compile with several of the recommended guidelines, see notes for explanations and the Memote report for other issues. Additionally, the README file associated with this model is sparser than ideal due to the lack of documentation associated with the original curation efforts.

Model Recognized naming convention o iPfal19: i + species indicator + iteration identifier model metadata (organism name, curation history, genome, authors, etc.) ORFs were called manually using proteomics and RNASeq data and compiled on the malaria parasite database; thus, no NCBI/refseq IDs (etc.) would accurately represent the genome used

Metabolite 100% of metabolites have a human readable, descriptive name 94% of metabolites have a charge and chemical formulaa 63% of metabolites have InChI strings, although because the pH for each subcellular compartment is known, these strings might not represent the appropriate species (i.e., protonation status)b 100% of metabolites have an ID from BiGG (100% have BiGG‐like IDs)c 91, 64.4, 68.7, 75, 47.5, and 68.4% of metabolites also have an ID from MetaNetX, KEGG Compound, ChEBI, ModelSEED, HMDb, or MetaCyc, respectively

Biochemical reaction 100% of reactions have a human readable, descriptive name 100% of reactions have a reaction formula 60.6% of reactions have an ID from BiGG (100% have BiGG‐like IDs)c 58.6, 20.7, 0, and 23.3% of metabolic reactions have an ID from MetaNetX, KEGG Reaction, ModelSEED, and MetaCyc, respectively 27.6% of reactions have an EC Numberb , d

Gene Gene IDs use PlasmoDB gene nomenclature, consistent with malaria field. These IDs map to a genomic location, DNA sequence, and protein sequence on PlasmoDB.org 71.1% of reactions have GPRs

Some metabolites do not have a charge or formula associated with them including metabolites representing host or parasite proteins. If many different proteins can participate in a reaction, the reaction contains a generic reactant to represent all of the possible protein reactants. Of the 6% problematic metabolites, nearly all are proteins or aggregate metabolites.

^aNot all BiGG IDs are mapped to InChI strings, EC numbers, or other useful identifiers (e.g., https://www.metanetx.org/chem_info/MNXM4217). This interferes with some Memote functionality, such as identifying duplicate reactions.

^bBiGG‐like IDs are proposed new BiGG IDs consistent with the general naming approach in BiGG. For example, new IDs (pheme_fv, pheme_ap) have been created for protoheme corresponding to protoheme located in parasite‐specific compartments, the food vacuole and apicoplast, respectively, consistent with existing BiGG IDs for protoheme (http://bigg.ucsd.edu/universal/metabolites/pheme). New reactions are created when the existing BiGG reaction occurs in only one compartment but should be present elsewhere in iPfal19. For example, “PLIPA2A120pp” is a periplasmic reaction in BiGG but occurs in the cytoplasm of P. falciparum; “pp” is the suffix used to denote the periplasm so the new cytoplasmic version is named “PLIPA2A120.” Additionally, new aggregate reactions (i.e., relevant pseudoreactions) were created and named intuitively (e.g., lipid1, lipid2).

^cTransporters, exchange reactions, and aggregate reactions should not have an EC number and these make up 35% of all reactions.