Making quality count in biological data sources

We propose an extension to the semistructured data model that captures and integrates information about the quality of the stored data. Specifically, we describe the main challenges involved in measuring and representing data quality, and how we addressed them. These challenges include extending an existing data model to include quality metadata, identifying useful quality measures, and devising a way to compute and update the value of the quality measures as data is queried and updated. Although our approach can be generalized to various other domains, it is currently aimed at describing the quality of biological data sources. We illustrate the benefits of our model using several examples from biological databases.