Video Tip of the Week: Canary Database for sentinels of human health

Recently we attended the Medical Library Association conference (#MLAnet13). Librarians are working so hard to wrangle information into usable forms, and to generate new connections among data types to reveal new information and leads for further studies. I ♥ librarians. In one of the sessions I attended on Medical Informatics, I heard several great talks. One speaker talked about a resource that connected types of information that hasn’t been well connected before: animal health + human health.

Sure–you think–we have information on cancer in mice. Rats. Disease models, etc. But this was different. Peter Rabinowitz was talking about connecting a different kind of animal information–data about non-model organism species that are part of the environment that we share. He and his team have been organizing that data into the Canary Database.

You might be aware of this already in some ways. Deer ticks and lyme disease. Horses and eastern equine encephalitis. Birds in China and #H7N9 flu. Agricultural animal transmission of antibiotic-resistant bacteria. But there are numerous other types of possible animal health consequences that could provide awareness and actionable information for environmental or human health interventions that would provide real benefits, or at least offer new leads to investigate.

Here you can watch a quick overview of the Canary Database:

This database has been around for some time, collecting this type of information. They have scoured numerous sources, including Medline/PubMed, CAB Abstracts, and Agricola. They curate this, and organize it in various ways to make it searchable and accessible. But it’s not easy to find the right papers, as we found out during the session at MLA.

In his talk Peter provided stories of the challenges of this work. There is no MeSH term for sentinel species types of data. So they had to text mine to find appropriate papers, and then filter through large amounts of work that wasn’t their target. For example–they were interested in companion animals that might have asthma symptoms. Cats in a home suffering from asthma might offer clues to interventions that could help humans who are affected. But a search for cats + asthma will provide lots of papers where the cats are possibly causing the breathing problems–and not the kind of work they wanted to feature for their database.

One way genomics folks could help them out would be to flag literature on this issue for them. Species research communities read a lot of papers, and if they find good examples of sentinel species it would help the Canary Database team out. Commenters at the talk also suggested contact with the MeSH folks to ask for a new category.

I can see the idea of this type of connection being really interesting and useful, though. And as we sequence more species, we could look for similar responses in a whole new range of organisms–and maybe identify new models of human illness that we hadn’t realized before. Or we can learn from the related genes and pathways. So in addition to the warnings of environmental hazards, we can study responses and interventions that would be helpful too.

So check out Canary Database, and learn about animals as sentinels for potential environmental and health warnings.

Quick links:



Rabinowitz, P., Scotch, M., & Conti, L. (2010). Animals as Sentinels: Using Comparative Medicine To Move Beyond the Laboratory ILAR Journal, 51 (3), 262-267 DOI: 10.1093/ilar.51.3.262

Rabinowitz, P., Gordon, Z., Holmes, R., Taylor, B., Wilcox, M., Chudnov, D., Nadkarni, P., & Dein, F. (2005). Animals as Sentinels of Human Environmental Health Hazards: An Evidence-Based Analysis EcoHealth, 2 (1), 26-37 DOI: 10.1007/s10393-004-0151-1

Rabinowitz, P., Cullen, M., & Lake, H. (1999). Wildlife as sentinels for human health hazards: a review of study designs Journal of Environmental Medicine, 1 (4), 217-223 DOI: 10.1002/jem.33