The value of currently available services and data resides not only in the information and knowledge they bring, but also in how they can be effectively integrated to infer new knowledge. With the growth in size of such data, integration and search become computational intensive tasks; planning and optimization of the way in which joined combinations are extracted is required to obtain results and get them efficiently. This paper discusses the challenging aspects of querying heterogeneous data sources and efficiently integrating the retrieved results, also effectively taking into account retrieved partial ranked values. The solutions adopted in the new Search Computing (SeCo) infrastructure, more precisely the server-side aspects of the framework, are illustrated in a Life Science scenario, where they allow discovering answers to complex biomedical questions.