Research on the progression of Alzheimer’smay 27, 2019
Project BReIN links research from a variety of professional fields
Five people contract dementia every hour. Making a diagnosis takes a long time, 14 months on average. With people under 65, this can even take four years. There is no cure yet, but there are medicines that can keep the symptoms in check. At Brightlands Maastricht Health Campus, people are looking for ways to speed up diagnoses and provide more insight into the progression of the disease. Together with companies based at the Brightlands Smart Services Campus in Heerlen, Kleinjans is going to look for ways to accelerate the computing power of data. The ultimate goal is to find a medicine.
Jos Kleinjans is a toxicologist affiliated with Maastricht University and is closely involved in the Brightlands e-infrastructure for Neurohealth (BReiN), the new name for the institute now that it will be studying the progression of Alzheimer’s disease. “We know that around ten percent of the patients have a hereditary form, a gene that is responsible for the disease. So how does it develop in the other 90 percent of patients? Scientists believe that heavy metals, certain pesticides and fats have an influence on how Alzheimer’s develops. We are going to look for markers in molecular processes that influence the disease, and we can use this information to look for better medicines,” Kleinjans says.
To find these markers, BReIN will first have to set up an extensive data infrastructure. In this infrastructure, existing platforms are linked to each other. These platforms include information from Scannexus, where MRIs are used to perform brain research among other things, and data from M4I, a research institute at Maastricht University where researchers study cell processes at the molecular level. “We are also setting up a new platform for genomic data with information on the structure and function of DNA,” Kleinjans explains.
By bundling together all of this different data, researchers hope to gain better insight into how Alzheimer’s develops in the brain. Kleinjans: “These days, every field is trying to find answers based on their own expertise but we take a different approach. We ask everyone to step outside of their discipline. In addition to brain samples from deceased patients and blood samples from live patients in different stages of the disease, we use cultured stem cell models with the same traits seen in Alzheimer’s patients. We work with MERLN here at the university on these activities. Useful information may also be obtained from meta data, or information on the lifestyles of deceased patients. This bundling of data is important because it gives you a more complete picture.”
All told, this involves an incredible amount of research data and a lot of computing power is required to derive anything useful from all of this data. Kleinjans: “This computing power isn’t available in the Netherlands. We would need a direct connection to one of the two supercomputers in the German city of Jülich; these systems are ranked number 26 and 44 on the list of the 500 most powerful computers in the world.”
Kleinjans is also going to look at ways to optimally utilize this computing power together with various companies based at the campus in Heerlen. Kleinjans refers to this as querying the data. “This might be an algorithm, a neural network or a simpler form of machine learning. You can even call it artificial intelligence as the hype prescribes. However, based on all the data from the experiments with stem cells, analysis of samples from deceased patients and from living patients in an early phase of the disease, we can partner with companies at the Campus to develop models that tell us something about the influence of aluminum on the brain in relation to Alzheimer’s, for example.”
How do you know if such a machine-learning model is accurate? Kleinjans: “This involves something called ‘blackboxing’, a situation in which scientists are no longer able to follow what a system comes up with. It’s not the end of the world as long as we know that the result is reliable. This is why experimental models must always be tested against practical ones, for example by using a control database for which the outcome is already known.”
Kleinjans has seen his professional field change with the arrival of genomics data, advanced analysis methods and more computing power. “The field has acquired more depth, also in a literal sense because we are capable of looking deeper and more accurately into cells,” says Kleinjans.
In spite of the fact that the Maastricht professor sees the advantages of better technology, we shouldn’t assume that this solves everything. “We shouldn’t be naive and just try everything because we can.” Kleinjans refers to a recent case of a Chinese scientist who modified an embryo through gene editing. “Suppose we find out which molecular process influences Alzheimer’s and can modify this through gene editing, how can you be sure that this procedure will only affect that one specific area? This is very difficult to ascertain, and will undoubtedly remain a topic of debate for years to come even though it is scientifically possible.”
Setting up this genomic data platform is being accompanied by a simultaneous expansion of the activities. “We are already working on this with the universities in Liège and Luxembourg among others. We want to be able to put the Netherlands on the map in this regard since this comes with huge opportunities. It would of course be fantastic if we find information in the genomic data that enables us to diagnose Alzheimer’s using a simple blood sample.” We are a long way from this, Kleinjans emphasizes.
It’s important to first properly set up all of the data structures. Kleinjans: “Once this is all underway, we can then look at whether or not data and any necessary analytical methods can be used as a service by the commercial parties at the Brightlands Smart Services Campus. But first we want to gain a better understanding of which environmental factors have an influence on Alzheimer’s so that we can then look for a better medicine. We also hope to find genomic information that improves the diagnosis. If this research method proves successful, we can apply it to other diseases, such as cancer. This can really change health care.”
(Source: This article is an abridged version of a recent article by Innovation Origins, an independent news platform and partner of Brightlands: www.innovationorigins.com.)