Reaction confidence scores

Every C. auto reaction in Genopsis has a confidence score from one to five. This is a measure of our confidence that a metabolic reaction is active in vivo.

Click the table below to see details of each level of confidence.

You must include justification in the 'confidence score evidence' field when assigning a confidence score.

This system is described in Daniell ( 2016):

A confidence scoring procedure was proposed by Thiele and Palsson ( 2010) but has seldom been used for published metabolic reconstructions. Confidence scoring is important as it allows the curator and user of the model to assess rapidly whether reactions are likely to be true. This information is especially useful when encountering a discrepancy between model prediction and experimental data. It is likely that there are reactions present in the reconstruction that would not be observed in vivo under any environmental condition. Gene-based homology evidence does not imply that these genes are expressed and translated to functioning enzymes. Furthermore, the RAST, SEED and Pathway Tools annotation engines seek to resolve false negatives through pathway-level prediction, further contributing to the presence of false-positive reactions. Finally, it is likely that there are reactions in the reconstruction with incorrect cofactors, directionality and stoichiometry. Our reaction confidence scoring system is described by and adapted from one developed by Thiele & Palsson (2010).