Power to the people: Citizen science meets precision medicine for rare disease

At this recent GoldLab Symposium presentation in Colorado, parent Matt Might shows how it’s done.

People credit rapid next-generation gene sequencing for the increased pace of medical discovery. But patients and their families—especially those with rare or undiagnosed conditions—are emerging as the true engines of precision medicine. Racing against the clock to save their children, parents are building databanks, connecting scientific dots and fueling therapeutic advances that could otherwise take a decade or more to happen.

“There’s a culture shift,” said Isaac Kohane, MD, PhD, chair of Harvard Medical School’s Department of Biomedical Informatics (DBMI), which hosted a conference titled Precision Medicine 2015: Patient Driven in late June. “A culture shift where patients feel empowered morally and intellectually to lead in precision medicine research and delivery.”

These patients are increasingly impatient. Impatient with Institutional Review Board (IRB) restrictions on sharing patient data across institutions or linking biological samples with clinical data. Impatient with research protocols that don’t give back their children’s data. Impatient with having to pay to get their children’s medical records and being locked out of those records once a child turns 14.

“That’s a problem when your daughter has intellectual disability and autism and is nonverbal,” said Megan O’Boyle, whose daughter Shannon, 13, has Phelan McDermid Syndrome, a rare condition caused by deletions of 22q13 or mutations of the SHANK3 gene.

Science for the people

O’Boyle and a growing number of parents are setting up their own studies, collecting blood and skin cells at family conferences, sharing data and wielding Google, social media and digital marketing as recruitment tools.

Matt and Bertrand Might
Matt and Bertrand Might

Conference speaker Matt Might, whose son Bertrand has NGLY1 deficiency, a previously unknown syndrome causing developmental delay, muscle weakness, seizures and an inability to make tears, broke the process down:

Step 1: Sequence your child’s exome to find the mutation. Bertrand’s was found in 2010 by Duke University,

Step 2: Interpret: What does the mutation do? Bertrand’s NGLY1 mutation causes loss of an enzyme that breaks down misfolded proteins; buildup of these proteins is thought to be the cause of his disease.

Step 3: Do science. The Mights worked with various scientists to find ways to counter the effects of the NGLY1 mutation, offering ideas from their reading. “Science becomes medicine, and science becomes action,” Might asserted.

Step 3b: Find a community. At exac.broadinstitute.org you can get an estimate of the number of patients with a given mutation. With a community, you can compare notes on symptoms, progress, test results and the results of any interventions.

In Bertrand’s case, though, there were no other patients to be found. So Might turned to social media. He wrote a blog post called Hunting down my son’s killer, deliberately crafting it to go viral and be easily found on Google. “We essentially applied search engine optimization,” he said.

The post was picked up by Gizmodo, Reddit, Hacker News and other blogs, and the family’s story was told last summer in the New Yorker. All this led patients to him. “If they were holding a diagnostic slip saying NGLY1, they were going to find us,” Might said.

In this post on internet-driven patient-finding, Might teaches how to jump to the top of Google searches, using marketing tools like AdWords to buy gene names, and by editing Wikipedia entries about the gene or disease. “Wikipedia is our secret weapon for finding patients—it has become the world’s database,” he said. “You can skip to the top of search results if you edit an entry.”

The NGLY1 community now has at least 37 patients.

Connecting the dots

Self-education is a common theme in rare disease: Sonia Vallabh and her husband Eric Minikel, who spoke later at the conference, went so far as to leave their respective fields of law and city planning to enter a PhD biology program at Harvard Medical School. They plan to devote their research careers to the prion disorder that killed Vallabh’s mother in 2010—and that Vallabh is destined to inherit.

Might, a computer scientist, had no prior biology training beyond two months in high school. He’s made up for it through relentless reading. “I didn’t even know what a chromosome was when all this started,” he said.

The Mights got Bertrand into an NIH Natural History study where he and other NGLY1 patients underwent a battery of “deep phenotyping” tests. These yielded many interesting leads, including a biomarker for the disease: hypoglycosylation, or an impaired ability to add sugar chains to proteins and lipids. The Mights used Google to find Hudson Freeze, PhD, an expert in glycosylation disorders, as well as an assay for NGLY1, which they sent to Freeze to screen for agents that might reverse deglycosylation activity. They learned that human NGLY1 protein was already available for sale, but also that the protein by itself might not constitute a cure, and that adding something called TAT peptide might help.

Eventually, through his investigations and reading, Might learned that when misfolded proteins accumulate in NGLY1 deficiency, they lock up a particular sugar, making it unavailable to the body. Suddenly, he had a plausible therapeutic strategy, one he could test immediately.

“I found a supplement on Amazon and took it myself—because in my house, I am the FDA,” he said. Feeling no ill effects, he gave the supplement to Bertrand. A few days later, his son cried his very first tear, and after three months, his nighttime seizures stopped.

Unlocking the data

Megan O’Boyle has been operating similarly. As a private citizen, she has been able to sidestep some of the regulatory constraints faced by academic researchers forming research registries and biorepositories. “I’m a stay-at-home mom and no one can fire me—I’m not going to lose any funding,” she said.

On the contrary, the nonprofit Patient-Centered Outcomes Research Institute (PCORI) has funded her to lead the Phelan-McDermid Syndrome Data Network. Its objective is to collect all available patient data “to make meaningful, well-annotated clinical data available to researchers and to share insights with members of the PCORI network.”

To ensure that those data truly are meaningful, O’Boyle has fought to help parents access their children’s medical records. “I don’t want these parents guessing when asked when their child walked, talked and had their first seizure,” she said.

Paul Avillach, MD, PhD, part of the DBMI and the Boston Children’s Hospital Computational Health Informatics Program (CHIP), works with O’Boyle to integrate all the information collected from parents, using natural language processing to dig out common threads in their narratives and search terms such as “epilepsy” to create aggregate profiles. “The idea is to unlock all the data buried in those clinical notes,” he said. “The whole idea is being able to touch the data, to generate hypotheses.”

Moving the needle

A variety of developments are facilitating data collection, such as direct-to-consumer blood testing and phlebotomists that can go to patients’ homes for blood draws. In a crowd-sourcing challenge organized by DBMI and Boston Children’s Hospital, called CLARITY Undiagnosed, research teams from around the world will soon compete to solve five families’ medical mysteries.

And on the analysis end, Might set up a fascinating “crowd screening” project called Mark2Cure, which sends volunteers scientific abstracts to annotate, highlighting relevant keywords and phrases to help researchers make connections. “They’re making the literature accessible to deep bioinformatics,” Might explained.

But connecting the dots would be a lot easier if the voluminous data already collected by the health care system could be readily shared and analyzed. The larger vision is to create an “information commons” bringing together patient genomes, epigenomes, microbiomes, “expose-omes” and other kinds of patient data. Conference speakers discussed the many institutional barriers—IRBs, health care IT systems that don’t talk to each other and research rules that some see as paternalistic, such as returning only “clinically actionable” genomic sequencing results.

“The phrase ‘not actionable’ does not apply anymore,” asserted Might.

Patient privacy and informed consent remain areas of uncertainty, concern and active debate in the bioethics community. Potential solutions to these issues exist—they just need a push to get implemented. And that’s where patients could wield the most power.

“It’s patient advocates going to Congress that are going to move the needle,” said Kohane. “Institutions—for a variety of reasons—are scared and we have to help them get unscared.”

This year’s Boston Children’s Hospital’s Global Pediatric Innovation Summit + Awards (November 9-10, 2015) is themed around Precision Pediatrics. For more information and to register, visit takingontomorrow.org.