Genome-scale model reconstruction protocol

Metabolic network reconstruction

Pipeline Overview

The metabolic annotation of protein sequences was performed with the E2P2 software (version 3.0) which is used to build Plant Metabolic Network Pathway Databases . To be more stringent when inferring homology between proteins, we decreased the Blast e-value cutoff in the E2PE code from 1e-2 to 1e-5. We used the pathway-tools software to infer biochemical reactions and metabolic pathways from the protein annotations (pathway-prediction-score-cutoff=0.3) and create a Pathway-Genome Database (PGDB) which can be browsed at this url. A MetExplore instance was also created to facilitate the visualization of the metabolic network and the mapping of omics data and is available at this url. .

Metabolic network curation

First, we removed 26 metabolic pathways from the initial set of 501 automatically inferred metabolic pathways. Then, we created a superpathway of sunflower’s oil metabolism on the basis of the main composition of the oil (refs) in the sunflower by merging 16 pathways:
  • superpathway of fatty acid biosynthesis II (plant)
  • fatty acid biosynthesis initiation I
  • linoleate biosynthesis I (plants)
  • oleate biosynthesis I (plants)
  • stearate biosynthesis II (bacteria and plants)
  • palmitate biosynthesis II (bacteria and plants)
  • fatty acid elongation – saturated
  • Vitamin E biosynthesis
  • diacylglycerol biosynthesis (PUFA enrichment in oilseed)
  • plant sterol biosynthesis
  • diacylglycerol and triacylglycerol biosynthesis
  • fatty acid biosynthesis (plant mitochondria)
  • gamma-linolenate biosynthesis I (plants)
  • long-chain fatty acid activation
  • monoacylglycerol metabolism (yeast)
  • very long chain fatty acid biosynthesis II
The oil biosynthesis pathway thus created involves 125 reactions, 160 metabolites and 429 genes. We checked also the links between genes and reactions by browsing literature about 40 metabolic genes (essentially identified in the oil metabolism).