Q&A with Rakesh Nagarajan
Siteman Cancer Center, Washington University
Rakesh Nagarajan is Assistant Professor, Department of Pathology and Immunology, Washington University; Co-Director, Co-director of the Siteman Cancer Center Bio-informatics Corps; and Director, Center for Biomedical Informatics. His perspective as an early adopter shows how caBIG® tools evolve over time to meet the most pressing needs of a research institution. Here, he shares his insights about how caBIG® addresses research and workflow challenges, as well as his vision for a system at Washington University where clinical studies, biospecimens and microarray data all compatible in a caBIG® fashion.
What are the biggest research challenges faced by your institution?
The two big challenges for us were very much related: data management and data sharing. The first problem of inefficient data management was related to the fact that there weren't good informatics to manage biospecimens at large banks. Most of the specimen information was stored on a local laptop or an excel spreadsheet. At Siteman Cancer Center, our main cancer bank had an Access database.
The data were highly localized and because the data were in different locations, investigators didn't know what kind of information could be accessed. The process for finding specimens was to call the person running the bank and put in an inquiry and many times that person wouldn't know if they had a certain sample or not.
Both data management and sharing were bottlenecks for research that are starting to be resolved with caTissue, which has been deployed here for almost two years.
What specific tools and technologies, has your organization developed or adopted?
We've led a lot of the activities in caTissue, which began as a single bank informatics system. Now, we are able to have multiple informatics systems using caTissue, and this is a major feature that is hugely beneficial. We now have four major banks using caTissue: one is our biggest cancer bank and the other three are non-cancer banks.
We also made the jump from caTissue Core to caTissue Suite, which offered us even more benefits. caTissue Core only held core data elements for specimen management, with very little annotation. Translational researchers want that additional information to determine which specimens to request.
Now, through the caTissue Suite, we can record annotated information right along with the tissue. On the other side, we have an interface that, with appropriate training, researchers can query more effectively and, therefore get the research approved must faster.
How did you work with the end-users at your organization to ensure that caBIG® would meet their needs?
On the management side, the director of our cancer bank was the "grandfather of caTissue" (Mark Watson). He and I are investigators in our own right, so we can wear that hat as we are considering what functions need to be included.
In the last 18 months, we have held direct conversations with our tissue bankers - they talk to us and to the development team to tell them about issues and products they wish they had. These are actual bankers who are entering data on a daily basis and then giving us regular feedback. Now, with the addition of new banks, our banker pool is growing and the feedback is growing.
There are a couple of scientists that we've trained to query on the system, and together we have identified usability issues that we'll address in the future. We created an interface that was robust but not as intuitive as a Google text box so we are exploring ways to improve the user experience.
Have caBIG® tools or technologies changed any of the clinical or research operations at your organization?
On biospecimen collection, we are starting to see changes in behaviors due to the web-enabled access to caTissue. For example, coordinators who are collecting specimens will enter preliminary data right then instead of waiting until the specimen comes to the bank and having the banker enter it. That leads to greater efficiency, and allows our researchers to focus on science, rather than process.
Do you have plans to adapt or adopt more caBIG® tools, or use your existing resources more broadly?
We have adopted caArray here and we're only using it to publish our microarray data set. We have a legacy array system that we've used to manage array data across the system and we would like to migrate that system into caArray directly.
This kind of migration would have significant cost implications but we envision a day when we will have integration of clinical studies, biospecimens and microarray data all compatible in a caBIG® fashion.
