High-quality computer infrastructures and resources are being used to help advance Alzheimer’s Disease (AD) research. Rob Quick — associate director of the Cyberinfrastructure Integration Research Center (CIRC) at Indiana University and presenter at the Gateways 2020 — is featured in this Science Node article discussing the disease, the field’s data distribution challenges, and the resources that are providing new hope for a breakthrough.
In the United States, as in many nations, society is aging.
This global trend toward aging, evident since the 1950s, reached a watershed mark in 2018 when the number of people above 65 surpassed that of young children for the first time in history. Of the over-65s, an average of one in 10 has Alzheimer’s Disease (AD), a ratio that increases to one in three at 85.
Policy and institutions devoted to aging research have proliferated in parallel with this demographic shift; yet a treatment that cures, or even slows, the biologically-mysterious AD has remained elusive.
To help overcome this stalemate, researchers have turned to today’s high-quality computer infrastructures and resources. The IUSM-Purdue TaRget Enablement to Accelerate Therapy Development for Alzheimer’s Disease (TREAT-AD) Center is leading an effort to make state-of-the-art AD data, computational tools, and methods broadly and rapidly available to AD research communities via a free gateway interface.
Kun Huang, director of TREAT-AD’s Center for Computational Biology Bioinformatics, and Rob Quick — associate director of the Cyberinfrastructure Integration Research Center (CIRC) at Indiana University and presenter at the Gateways 2020 — spoke to us about the insidious and puzzling disease, the field’s data distribution challenges, and the resources that are providing new hope for a breakthrough.
The project is designed to facilitate the discovery of new high-confidence targets — molecules associated with the disease and targeted during drug treatment — and ultimately, by effect, accelerate drug discovery and treatment.
The data, inherently heterogenous and big, draws from extant databases on genomic mutations and sequences, biological structure, chemical informatics, and cell and biochemical assays — along with TREAT-AD’s recent databases on neuroinflammation mechanisms, which are the current focus in the field’s shifting tide of theory on AD etiology.
And, in turn, the infrastructure is equally heterogenous: It synthesizes existing and novel, TREAT-AD-developed infrastructures as well as highly-used open-source and commercial tools (like CDD Vault). This heterogeneity made developing an effective infrastructure — with a complex supercomputing backend and simple, usable frontend — especially difficult for the researchers:
“The challenge is, on the one hand, we want to develop a nice user interface so everyone in the community can use it. On the other hand, we need supercomputers on the backend, because proteomics tools, and some of the imaging tools we’re developing right now, are computationally heavy,” says Huang. “You need either large clusters or GPUs to run the AI and informatics processing.”
To meet the interface needs of unique users, from structural biologists to computational chemists, the researchers developed a specialized, flexible gateway that allows the addition of new tools — a feature that will be especially useful in the future as new research techniques become feasible.
Huang says that the infrastructure core also promises utility in translation to other medical research fields like cancer, where its applications are already being studied.
Despite the disease’s prevalence and seriousness, pharmaceutical companies have recently scaled back their drug discovery operations after a series of costly failures.
Scientists know, for example, that “plaque” and protein tangles, known as “tau,” accumulate in the brains of individuals with AD. But they do not know the role the proteins play in brain degeneration (or whether or not they might be merely an effect of some other cause). In fact, “There is actually a small group of patients who can have very heavy tauopathy, but very mild symptoms,” says Huang.
To improve knowledge on the matter, the National Institute on Aging (NIA) launched multiple data collection initiatives between 2014 and 2016. Carried out by the nation’s first governmentally-funded, multi-institution (and -national) teams, the initiatives tackled modeling in mice and AD-resilience analysis in individuals.
The mouse models — developed in a collaboration between Indiana University, the Jackson Laboratory, Sage Bionetworks, and The University of California, Irvine — marked a significant shift toward more accurate AD modeling in organisms, the lack of which has been widely cited as stunting the field for decades.
The broad accessibility of this cutting-edge data, along with the massive amounts of data collected from other NIA-funded programs, marks another shift: Throughout its history, intellectual property barriers to research have slowed the field’s advancement. For example, engineering a certain AD mutation in mice was, for a time, the exclusive and patented right of a single institution, the Alzheimer’s Institute in America (AIA).
While the task of uncovering AD’s mysteries remains challenging, the infrastructure’s data accessibility and novel technologies will give researchers new ways to investigate, providing new hope for treatments and cures.
“When I think about this project, I think that someday, somewhere someone might not have to suffer as badly from Alzheimer’s Disease as they would today. That’s kind of lofty, but it’s why we need continued biomedical research in this area.” says Quick.