David & Nielsen 2005 — source map, not a second engine¶
Working note from a photographed copy of:
David, H. & Nielsen, J. (2005). Modelling of fungal metabolism. In Vaidyanathan, S., Harrigan, G.G. & Goodacre, R. (eds.), Metabolome Analyses: Strategies for Systems Biology, ch. 12, pp. 195–214. Springer, Boston, MA. doi:10.1007/0-387-25240-1_12
Companion to gem-primer.md. Same altitude: genome-scale reconstruction and flux balance. Not a FoodPacket walk.
What the chapter is doing¶
It treats a GEM as an iterative loop (their Figure 1, p. 197):
- in vivo real network ↔ reconstructed network (in silico)
- reconstructed network → simulated behaviour
- simulated behaviour ↔ measured behaviour
- data integration in the middle
That loop is constraint-based modeling: build S, add bounds, choose an objective (often maximal growth), let FBA fill fluxes, then go back to the bench. Covert et al. (2001) and the later S. cerevisiae reconstruction (Förster / Nielsen, 2003) sit on the same line.
The paragraph under Table 2 is the sentence that matches this package: sequence homology is not function. The same annotated EC on two genomes does not license the same physiological flux. Presence of a reaction is a separate seat from how much carbon moves.
Table 2 — printed URLs vs 2026¶
| Printed name | Printed URL | 2026 |
|---|---|---|
| EcoCyc | http://ecocyc.org | Live. E. coli K-12 literature GEM. EcoCyc.org; still versioned (29.x). Free. |
| MetaCyc | http://metacyc.org | Live. Experimentally curated multi-organism pathways inside BioCyc. |
| MPW | ergo.integratedgenomics.com/MPW | Public tool gone (Integrated Genomics / ERGO era). |
| KEGG | genome.ad.jp/kegg | Live as kegg.jp. Reference maps, different curation model than EcoCyc. |
| WIT | wit.mcs.anl.gov | Service gone; lineage continues in SEED / ModelSEED. |
| Biology WorkBench | workbench.sdsc.edu | SDSC service gone. |
| EMP | empproject.com | Historical enzyme/pathway archive; not a reconstruction engine. |
Use EcoCyc / MetaCyc / KEGG as citation targets for pathway identity on LAW cards. Do not vendor them. Do not treat a KEGG map id as a measured meal flux.
What this package does not take from the chapter¶
- An FBA solver or a second
simulate_mealthat emits flux mmol. - COBRApy in the kernel (GPL-2.0 vs Apache-2.0; already listed adjacent in related-projects.md).
- A CI gate that fails a PR when a GEM growth rate is off 5%.
- Filling OPEN amounts from an objective function.
A GEM answers which fluxes are feasible under an assumed objective. This package names digestion and teaching paths and leaves magnitudes empty unless a seat is declared. Complementary, not substitutable.