On quality assurance and assessment of biological datasets and related statistics

The complexity of modern biological database management systems indicates the need of integrated metadata repositories for harmonized and high-quality assured data processing. Such systems should allow for the derivation of specific producer-oriented indicators monitoring the quality of the final datasets and statistics provided to the end-users. In this paper, we offer a quality assurance and assessment framework for biological dataset management from both the producers' and users' perspective. In order to assist the producers in high-quality end-results, we consider the integration of a process-oriented data/metadata model enriched with quality declaration metadata, like quality indicators, for the entire process of dataset management. With the automatic manipulation of both data and "quality" metadata, we assure standardization of processes and error detection and reduction. Regarding the user assessment of final results, we discuss trade-offs among certain quality components (such as accuracy, timeliness, relevance, comparability, etc.) and offer indicative user-oriented quality indicators.