In molecular and cell biology, most of the data presented in published papers are not available in accessible formats that would allow for analysis and systematic mining. Here we present SourceData (http://sourcedata.embo.org), a platform that allows researchers and publishers to share scientific figures and, when available, the underlying source data in a way that is machine-readable and findable. SourceData has therefore developed tools to generate machine-readable descriptive metadata from figures in published manuscripts. Experimentally tested hypotheses are represented as directed relationships between standardized biological entities, which can be connected into a searchable data-oriented ′knowledge graph′. SourceData focuses on the core of scientific evidence - data presented in figures - and makes papers searchable based on their data content. By coupling data availability to improved discoverability, SourceData aims at establishing a self-reinforcing data ′ecosystem′ that bridges the conventional visual and narrative description of research findings with a machine-readable representation of data and hypotheses.

SourceData - a semantic platform for curating and searching figures