It would also be helpful to understand what metabolic reactions are being affected by mutational events. One of the included metabolic models for E. coli contains this information, so we need to decide how to best use it for analysis.
Some questions to think about:
For novel phenotypes, how we can add a chemical similarity to go from novel compound to something the DB has seen before? More generally, is there a way to introduce a proper phenotype ontology to help classify compounds/conditions at different levels of specificity?
Do we want to just use the metabolic models directly rather than adding them to the DB?
Do we want to store the protein-substrate and protein-ligand relationships somewhere?
Can we use RetroRules or another database to predict substrate promiscuity?
Can we use chemical group analysis to predict toxicity?
It would also be helpful to understand what metabolic reactions are being affected by mutational events. One of the included metabolic models for E. coli contains this information, so we need to decide how to best use it for analysis.
Some questions to think about: