AMAS-sb 0.0.4

AMAS (Automatic Model Annotation System)

AMAS: Automatic Model Annotation System

AMAS (Automatic Model Annotation System) predicts and recommends annotations of systems biology models. Current version focuses on species and reactions of models in SBML format, which can be tested using BioModels or BiGG models.

Overview

AMAS is a collection of methods to predict and recommend annotations for SBML model elements. Current version focuses on predicting species and reaction annotations of metabolic models, such as that can be found in BiGG and BioModels repositories. Algorithm uses CHEBI for species annotations, and Rhea for reaction annotations.

Example

First, class instance should be created using AMAS.recommender.Recommender with an existing SBML file (optional).

Next, to get recommendation for species, user can use the .getSpeciesAnnotation method.

When the pred_id argument is used, Recommender will search the model to find an available display name for prediction. Alternatively, user can use the pred_str argument to make a direct prediction, which does not need a pre-loaded model in the constructor. Below, 'S-adenosyl-L-methionine' is the display name of the species 'SAM' in the model file, so the result will be the same.

Result is a namedtuple 'Recommendation', with attributes including id, credibility, recommended CHEBI terms, and the urls of such terms.

Similarly, recommendation of a reaction can be also obtained using the .getReactionAnnotation method.

Recommendation of a reaction uses Rhea database.