Choice of annotation data significantly impacts automated reconstruction of genome-scale metabolic networks

Changing the source of reconstruction data substantially alters the contents and predictions of automated genome-scale metabolic models, with data source effects exceeding those from reconstruction method choice.

• Why it matters: Curated genome-scale metabolic models exist for only limited organisms, and automated draft models often lag; improving them requires understanding how underlying annotation data shapes outcomes.
• What they did: The study varies the reconstruction data sources and tests sensitivity to gene annotation accuracy, then evaluates how functional annotation from closely related organisms affects automated model reconstruction.
• The result: Model outputs are strongly dependent on reconstruction data selection and annotation accuracy, implying that careful choice of functional annotation—especially from closely related organisms—can materially improve draft model reliability.

AI summary of the abstract · may contain errors · not medical advice · read the paper.