The Biobank Dilemma

Killing two birds with one stone. I am currently preparing for a presentation for the Scientific Advisory Board of our local biobank network P2N and I thought that I wanted to share a few thoughts on biobanking and sample collection for epilepsy genetics that have troubled me for a while.

Left behind in the biobank world. Biobanking is a big thing these days and virtually every institute has established or at least played with the idea of establishing a biobank. I might be influenced by the fact that biobanking is currently very fashionable in Germany, but I think the move towards declaring large sample collections biobanks is an international phenomenon. Biobanks have one thing going for them: their huge size. Quality control issues for sample collections, standardized pathways for extraction, sophisticated protocols for data protection etc. only make sense if you are dealing with thousands, if not tens of thousands samples. These biobanks are considered the “gold of the 21st century”. The bad message is that this system is not very useful for epilepsy genetics.

Phenotypes? Epilepsy genetics is dependent on making phenotypic information available, information that is often difficult to estimate beforehand. Imagine that we find a novel mutation in two patients with epilepsy. We would be interested if the epilepsies have similar features, something that is difficult to derive from a standardized phenotyping sheet. The most useful thing is in fact a good and detailed patient letter, which allows you to trace back the phenotypes. In most large-scale biobanks, patients are anonymized, which makes the acquisition of a patient letter impossible. The process of refining patient phenotypes has been named “iterative phenotyping”. Even though this phrase implies some algorithm-like approach that can be carried out systematically for each and every phenotype of choice, there is actually little discussion on how this iterative phenotyping can be performed. In fact, if you have a state-of-the-art biobank, this might be virtually impossible. In Kiel, we are very lucky to be an exception to this.

Diagnostics? The other end of the spectrum is diagnostic testing or clinical research closely affiliated with it. While the framework for clinical genetic testing is well established, it carries the inherent problem that is has to rely on established genetic tests. Novel diagnostic methods such as panel sequencing or some day exome sequencing may blur the boundary, but clinical genetic testing is very conservative by nature. You only report on findings that you are very certain about, the remainder is uncertain. And we work in epilepsy genetic research because we need to push this boundary, as we can explain so little.

The spectrum of patient proximity in DNA sample collections, reaching from diagnostic testing to large-scale anonymous biobanks. The dilemma for epilepsy genetics is the fact that we are stuck in the middle. We need large sample sizes, but also access to additional clinical information (iterative phenotyping). This balancing act between both well-defined extremes has difficulties attracting the necessary large-scale funding.

Stuck in the middle. The kind of sample collection we need for epilepsy genetics is located in the middle of both extremes. We need close contact with our research participants for iterative phenotyping while ensuring that we do not confuse research with diagnostic testing. There is a tendency by both on the patient and researcher side to mix both areas up. Two months ago, I received a call from a colleague asking us whether we could include one of his patients in our exome studies since the parents wanted to get exome sequencing as fast as possible. I tried to explain that we don’t promise time frames and don’t even promise inclusion of samples to participants and that it is important for us not to give patients the false impression that we are on the verge of solving the disease. As discussed in many previous posts, the chance of finding a potentially causative variant through exome sequencing is less than 10%, slightly exceeding the pick-up rate of a diagnostic array CGH. On the other hand, patients are keen to have research results returned, which is a strong motivator for participation in the first place. Therefore, genetic research particularly on severe epilepsies is always about trying to find the right balance.

Why would you fund a balancing act? Large biobanks, sometimes pejoratively referred to as “sample farms”, impress through the industry-like work flow and throughput, which has the capacity to attract large infrastructural grants. On the other extreme, diagnostic testing is covered by health insurance companies or sometimes by patients privately. In brief, there is funding on both ends of the spectrum. So what is left for our balancing act in the middle? What the field of epilepsy genetics needs is central resources for sample collection that are able to retrieve primary data from patients and eventually even re-recruit patients for research studies. This is nothing a single research group can handle and is currently nothing that funding bodies seem to be particularly interested in. As a response to the Waterloo Foundation meeting on Idiopathic Focal Epilepsies, we suggested generating such a resource through the funds of a foundation in parallel to the Simon Simplex Collection in autism research. If resources like this are created, the perceived balancing act may someday be a basic part of modern epilepsy genetic research.

Ingo Helbig

Child Neurology Fellow and epilepsy genetics researcher at the Children’s Hospital of Philadelphia (CHOP), USA and Department of Neuropediatrics, Kiel, Germany

Facebook Twitter