With the advancement of Web, novel types of scientific-related data and metadata are emerging from a growing number of various sources. Alongside traditional bibliographic data provided by digital libraries great amounts of social metadata (such as bookmarks, "reads", tags, comments and "likes") are created and accumulated by social networking services. We believe that these metadata can be fruitfully used for improving search and assessment of scientific knowledge. The individual sources of scientific metadata differ largely in their focus, functionality, data coverage and data quality, and are currently limited to their own databases and data types. We suggest that we can enhance the current individual services by integrating their data and metadata. In this paper we discuss the opportunities and challenges of such integration for the purpose of facilitating the discovery and evaluation of scientific knowledge, and present a framework for integration and warehousing of both bibliographic and social scientific metadata.