After Bio-IT 2015 in Boston two weeks ago, having recovered from the jet lag and too many late nights of discussions & networking (more on our own Pistoia Alliance meeting later), we were reflecting on what was different from previous years and what the current trends appeared to be.
It was clear that there are increasing trends towards cloud-based/SaaS approaches to informatics delivery, plus a steady stream of new vendors still entering the market of genomics and NGS analysis.
Equally, the talks were varied and the dense tracks meant that you had to be picky about the ones you attended otherwise you could miss the overall theme of each track. There was much discussion on learnings of best practice in the deployment of new capabilities and the chance to review pre-competitive groups in different parts of the workflow.
One thing that seemed relevant is that knowledge tools are maybe coming of age, and this is in part due to the drive and pressures we face in biomedical research with data from so many sources. It is also due to people describing their data better, and the tools to extract knowledge & inferences are getting better at answering some of the key questions.
From the behemoth of IBM’s Watson engine and its recent introduction to oncology decision support, to various search & text mining tools, all of these approaches will depend on rules on the definition of entities and the descriptive vocabularies and/or ontologies that are used.
We feel this fits in well with our own Ontologies Mapping project that is just getting going, and if you are either building your own knowledge engineering environment or seeking to use a commercial tool, you should be pleased that ontologies are there to help you.
Ontologies are often the unsung aspects of integration and knowledge management but they are nonetheless important and our goal is to help map the equivalence between ontologies that may have evolved from different groups. Being able to bring together disparate ontologies and querying them as one, while preserving the original context and implicit meaning behind the chosen annotations, will be a very useful result for anyone who has to work with research data.