Abstract
Persistency of genomic data brings many research challenges. Among them, alternatives to the commonly used relational database systems should be evaluated, to deal with large volumes of data, in particular writing operations and specific modelling issues. In this paper, we present a NoSQL approach to treat genomic data (Cassandra database system), and evaluate both persistency and I/O operations in experiments with real data. Our results show that this approach enables to store, access and manage genomic data with performance gains, when compared to relational database systems.