Strengthening digitalisation in preclinical research
"Our project will use text mining and artificial intelligence to systematically process the data from neuroscience publications and make those data available to scientists free of charge", says Benjamin Ineichen.
The process of developing treatments for neurological diseases such as Alzheimer's, multiple sclerosis or Parkinson's is extremely complex. While there has been some progress in recent decades, there have also been many setbacks. A major problem is that many approaches that look promising in animal studies fail in human clinical trials. In Alzheimer's research, for example, an estimated 99 percent of potential new treatments produce no effect in humans. This is a particular problem because neuroscience research is responsible for more than a quarter of all animal experiments in Switzerland, and there is a relatively high proportion of severity level 3 experiments (highest stress for the animals).
Various factors contribute to this situation, including the fact that animal studies are not always planned and conducted as well as they could be, often because a lack of data makes it unclear what the optimal method would have been in the first place.
An addition hurdle is the fact that animal studies generate so much scientific data each year that it is impossible to keep track of it all. For example, it is estimated that more than one million scientific articles are published each year in biomedicine alone.
"Scientific data overload is an urgent problem that will only get worse in the future. We need to find ways of making this mass of data available to researchers so they can take an evidence-based approach to planning animal experiments," says Benjamin Ineichen from the Center for Reproducible Science in Zurich, summing up the situation.
Ineichen's team therefore plans to create a database for neuroscientists: "Our project will use text mining and artificial intelligence to systematically process the data from neuroscience publications and make those data available to scientists free of charge and in a clear form." This will help researchers to better plan their animal experiments, and the process will evaluate several hundred thousand studies. "Our project is designed to further the digitalisation of preclinical research," says Ineichen.
ANIMONE: An in vivo data warehouse for neuroscience – biocuration of preclinical research to inform optimization and exploitation of preclinical studies