Abstract Recent advances in the automation of metabolic model reconstruction have led to the availability of draft-quality metabolic models (predicted reaction complements) for multiple bacterial species. These reaction complements can be considered as trait representations and can be used for ancestral state reconstruction to infer the most likely metabolic complements of common ancestors of all bacteria with generated metabolic models. We present here an ancestral state reconstruction for 141 extant bacteria and analyze the reaction gains and losses for these bacteria with respect to their lifestyles and pathogenic nature. A simulated annealing approach is used to look at coordinated metabolic gains and losses in two bacteria. The main losses of Onion yellows phytoplasma OY-M, an obligate intracellular pathogen, are shown (as expected) to be in cell wall biosynthesis. The metabolic gains made by Clostridium difficile CD196 in adapting to its current habitat in the human colon is also analyzed. Our analy..." /> rel="meta" type="application/atom+xml" href="http://dx.doi.org/10.1089%2Fcmb.2013.0079" /> rel="meta" type="application/rdf+json" href="http://dx.doi.org/10.1089%2Fcmb.2013.0079" /> rel="meta" type="application/unixref+xml" href="http://dx.doi.org/10.1089%2Fcmb.2013.0079" /> 展开▼