W08: Noctua: An all in one solution for ontology development and annotation

Title

Noctua: An all in one solution for ontology development and annotation

Workshop type

Tutorial

Organizer

Suzanna Lewis, Lawrence Berkeley National Laboratory

Translating biological data and information into actionable knowledge demands a biological resource containing not merely a large amount of data, but a fully integrated resource of contextualized interrelated facts. We have implemented an initial knowledge acquisition environment, “Noctua ”, that supports collaborative real-time editing of biological models, backed by the semantics of multiple integrative ontologies. In this workshop we will carry out three exercises to demonstrate and train users in accurately capturing information from published papers. The examples will include:

1. gene function and molecular pathways;
2. metagenomic environments; and
3. anatomical phenotypes associated with a Mendelian disease.

Small teams of 2-4 attendees will each independently carry out their own annotation for these three papers using the online interface to draw models of the mechanisms for each step of the biological process and how these are wired together. Following these independent annotation efforts the resulting models from each team will be compared and discussed. This hands-on work will demonstrate the collaborative environment, and how researchers can directly author computable descriptions (covering both ontological classes and specific instances) of causal biological models using a graphical interface.

Biologists of all stripes, whether developmental, cellular, physiological or ecological, are model builders: models that reflect their current understanding of how a biological system works. If we could equip machines with the ability to understand and compute over these models, then these machines would be able to help us carry out the task of analyzing and interpreting the exponentially growing amount of data generated by biological researchers. In order to achieve this goal, we need to be able to communicate these interconnected models to computers in a way they can understand. Humans communicate these models to one another by means of diagrams, but drawings cannot be read or comprehended by machines.  A machine cannot know what the labels on the boxes or the labels on the edges connecting things together mean. In bioinformatics, the standard way to add semantics is to use an ontology, which connects together a set of terms used to describe a domain in a graph. These wiring diagrams are represented in a standardized language (OWL) shared by both scientists and machines, allowing computers to gradually build up an internal model of biology, ultimately allowing us to use advanced computational techniques to probe for causes and treatments of disease and other natural perturbations. The motivation for this workshop stems from a desire to empower more biologists with the ability to draw their models in semantically transparent way, to build this community, and widen the discourse.