Presenters from Boston Children’s Hospital and Iceland’s deCODE genetics detailed the impact of sequence-based diagnosis of rare disease at WuXi NextCODE’s “Genomes for Breakfast” series at ASHG17.

My last post described work on rare diseases at Children’s Hospital of Fudan University (CHFU) in China, but at our recent “Genomes for Breakfast” series at ASHG17, we also heard about the impact of sequence-based diagnosis of rare disease from colleagues in Boston and Iceland.

From our longstanding partners at Boston Children’s Hospital (BCH), we heard a detailed discussion of how sequencing hundreds of people across numerous families, and the analysis of all that data together, was accelerating our understanding of one disease: nemaline myopathy. That was presented by Alan Beggs of BCH’s Division of Genetics and Genomics, and the Manton Center for Orphan Disease Research at BCH/Harvard Medical School.

A myopathy is a disorder affecting the skeletal muscles. Typical symptoms of congenital myopathies include early onset of hypotonia and weakness, which reflect distinctive pathologic changes in the muscle fibers. There are multiple subtypes of myopathy. Nemaline myopathy is the most common subtype. It can cause painful contortions (arthrogrypothis), make patients dependent on ventilators and wheelchairs or, in less acute forms, lead to mild weakness that can still impair mobility.

One of the biggest challenges in understanding these myopathies is that they are heterogeneous in every way—clinically, pathologically, and genetically. There are multiple genes associated with different subtypes, and they are sometimes shared but can also be completely distinct. Scientists have also found multiple patterns of inheritance even for a single gene.

To try to shed light on this genetic puzzle, Beggs and colleagues have been looking in detail at the particular mutations found in certain patients and mapping those against their symptoms and inheritance patterns. Recently, the researchers sequenced 857 patients with several subtypes, including just over 300 with nemaline myopathy. The BCH researchers are using a variety of sequencing approaches, including exome analysis, RNA sequencing, and whole-genome sequencing (WGS). This is necessary, they have found, because there are unusual variants that are difficult to identify with traditional tools.

Using this array of sequencing technologies and WuXi NextCODE’s clinical interpretation and case-control research tools, they were able to identify specific genetic variants that have already been associated with certain subtypes, identify new variants, and start to predict the clinical impact of each mutation or constellation of mutations. These tools are able to instantly draw upon a wealth of data from BCH, WuXi NextCODE’s knowledge base, and public databases.

The picture that emerges is complex: a large set of variants that sometimes overlap across multiple subtypes, but sometimes are just strongly associated with a single subtype. The NEB gene is of particular interest because it is so strongly associated with nemaline myopathy and quite a bit is known about its biology. Based on these studies, Beggs and his colleagues have also suggested a more accurate approach to diagnosing nemaline myopathy. All of these are steps that benefit patients and their care, and they provide the first step toward developing new and more effective ways of treating these conditions.

The third country represented at our second breakfast was Iceland. Patrick Sulem, who leads the clinical team at deCODE genetics in Reykjavik, presented. Like our collaborators in Shanghai and Boston, but armed with a truly unique set of resources and expertise, the deCODE scientists have made significant strides in better diagnosing rare diseases in children.

The deCODE database is unique in many ways and powered to uncover rare disease-causing variants. It includes the directly sequenced whole genomes of nearly 50,000 Icelanders and 10,000 others; imputed whole genome data on some 400,000 Icelanders; and SNP data from nearly a million people around the world. This gives deCODE allelic frequency data of unrivaled detail. As we have shown, this data can be helpful in diagnosing disease around the world—but when used in Iceland itself, it can point straight to pathogenic mutations and provide a map to wherever they lie in the population.

These strengths are based on some advantages of the population approach, some that others would like to replicate, others that are tough to match. Iceland’s population participates in genetics studies at a higher rate than that of any other country; Iceland has a long, strong tradition of preserving ancestry records and so has a nearly complete national genealogy for the modern era; and a national health system with a centralized record system. These ingredients have given deCODE the right data to find important variants in diseases that have baffled others. (For more details, read my post on Kari Stefansson’s headlining talk for our breakfast series.)

Based on deCODE’s work, it is now evident that whole-genome sequencing (WGS) can greatly improve diagnosis and clinical management of infants and children with hard-to-diagnose diseases. Like their peers at BCH and CHFU, researchers in Iceland have been able to use genomic screenings not just for better diagnoses—giving parents at least the comfort of knowing what’s wrong—but also, in some cases, they have been able to offer better guidance for the children’s treatment.

One such case was the result of the early application of our technology, before NextCODE had spun out of deCODE. It involved two sisters who had undergone a diagnostic odyssey of several years. With whole-genome sequence data from them and their parents, and with the ability to filter allelic frequency data in the context of different modes of inheritance, we were able to identify the culprit variant—a previously unknown variant causing Brown Vialetto Van Laere syndrome—in a matter of minutes. Because the variant was disrupting a riboflavin transporter gene, the diagnosis immediately suggested riboflavin therapy, a course of treatment that halted the progression of their disease.

Finally, it is important to note that the identification of rare disease variants is a promising avenue for feeding drug discovery—not just for the rare conditions themselves, but also potentially for much more common conditions, of which rare diseases can be extreme versions. Solving rare disease is a challenge for us all—indeed, it is a common challenge in the truest sense. The more diagnostics we do, the bigger our databases all over the world, and with the informatics and tools to mine all this data, the more benefits we can deliver to people around the world.

In just one year, Fudan Children’s Hospital and WuXi NextCODE have diagnosed 11,000 pediatric patients in China and created a program that rivals the largest labs in the U.S.

Children’s Hospital of Fudan University (CHFU) in Shanghai is widely considered China’s top pediatric hospital. The doctors there see almost 2.5 million patients annually.

One short year ago, WuXi NextCODE and Fudan launched sequence-based rare disease

testing at CHFU using WuXi NextCODE’s RareCODE test and backed by our knowledgebase and the collective expertise of both organizations. In this first year, an astonishing 11,000 patients received sophisticated genomic screening tests to help guide treatment for hard-to-diagnose, or rare, diseases. One-third of those patients got a precise diagnosis, matching the best rates anywhere in the world. In short, Fudan and WuXi NextCODE have, in just one year, effectively launched the field of sequence-based rare disease diagnostics in China and created a program that rivals the largest labs in the U.S.

We had the distinct pleasure of hearing directly from doctors handling these cases and scientists building the database supporting this collaboration at our recent ASHG breakfast, “Using NGS to diagnose rare disease—experiences from three continents.” Dr Lin Yang, MD, PhD, a clinician at CHFU’s National Children’s Medical Center, presented the hospital’s experience with this rapidly expanding new program.

The service was created thanks to the unique partnership established between the hospital and WuXi NextCODE. WuXi NextCODE contributes its know-how in clinical-grade genomic sequencing, massively scalable informatics, and RareCODE test, backed the most powerful interpretation tools and clinical genetics expertise available.

CHFU, meanwhile, brings to bear the services, knowledge, and skill of its pediatric specialists and the national center of excellence in pediatric medicine housed at the hospital. Notably, CHFU’s Institute for Pediatric Research had previously developed more than 100 tests for single-gene genetic diseases, established multidisciplinary teams of clinical experts to address rare disease, and is among the very first hospitals in China to adopt next-generation sequencing.

Armed with our IT, knowledgebase, and diagnostic tools, this pioneering collaboration has advanced a national center of excellence for diagnosis, treatment, and further medical genetic research. At its core are not just the expertise of both teams, but also a rapidly growing database of mutations causing rare diseases, which the team hopes to grow into the largest in China, and perhaps the world.

Just over 5.5% of babies in China are born with some type of evident syndrome or birth defect. There, as elsewhere, these can impact the skeleton, metabolism, nervous system, circulation, respiration, digestion, and more. These can also be very complex, with multiple phenotypes or overlapping disorders. Some of these are due to causes other than genetics. But a large proportion represent genetic syndromes, of which many are de novo or have never been seen, or at least written about, by other clinical groups.

CHFU started doing single-gene sequencing to help resolve such cases as early as 2010. By 2012 the hospital was also running array CGH, and in 2013 it launched a number of panel tests and an NGS data-analysis pipeline. This history of pioneering genetic analysis put the hospital at the forefront of medical genomics. And things really moved forward fast after CHFU created a joint molecular diagnostic laboratory with WuXi NextCODE.

The two groups confer weekly on difficult cases and, as of now, they have completed some 12,000 genome analyses in just one year, providing a diagnosis in 33% of cases. These include the smallest patients, from the neonatal intensive care unit (NICU)—more than 2,200 of whom received focused exome sequencing and analysis. Just over 13% of those infants received a diagnosis. This lower rate of diagnosis among newborns reflects the greater challenge of working with patients whose signs and symptoms are just appearing. But this number is rising and, as the NICU is a first-tier clinical setting, every diagnosis can be a lifesaver.

Parents and other family members are also often sequenced to determine if the mutations are passed down or have occurred spontaneously (i.e., are de novo). All of that data is incorporated into a database, helping to grow knowledge about the mutations that cause rare diseases.

While there is no specific treatment for most of the syndromes identified, there is an improving picture for a growing number. Patients may receive a lifesaving, or life-changing, treatment plan, or referral to specialists based on their anticipated future needs. Regardless, it is important for the family and doctors to understand as much as possible about what the problem is. It is also helpful for parents and relatives to know that there are potentially pathogenic mutations that run in the family.

At ASHG, our head of communications, Edward Farmer, sat down with Dr Yang, other scientists and physicians from CHFU, and our CSO, Jeff Gulcher, to talk about the growth of the WuXi NextCODE joint rare disease lab and some of its early successes. I’ll be posting Dr Farmer’s interview with them here in the days ahead, so be sure to check back and learn more about the launch of rare disease testing in China.

In an interview with WuXi AppTec’s WXpress news site, WuXi NextCODE CEO, Hannes Smarason, summarizes how genomics AI can make drug development better, faster, and cheaper.

How will WuXi NextCODE bring AI to the forefront of drug discovery and development? I mapped that out recently in an interview with WuXi AppTec’s WXpress news site. Below, I summarize some of the key points from the article.

Our advantage
The core of our strategy is to bring together three different things: Cutting-edge algorithms, domain expertise, and large data sets.

Executing on the strategy
Because of our history pioneering this field, we already have a wealth of samples as well as a very sophisticated and robust way of mining that data to help discover novel treatment pathways and for possibly re-purposing drugs. We are also continually developing new algorithms and other methods to distinguish between people who will respond and those who won’t when given a particular treatment.

The benefits of AI for drug development
Incorporating AI creates a much more data-driven rather than hypothesis-driven process. That improves the likelihood of identifying patterns and novel insights that may have been overlooked using conventional methods. That means finding the truly causal genes or pathways that drive disease. The goal is to have a more powerful starting point for the development of treatments.

Early validation
Our deep-learning algorithms have already been used to uncover a particular mechanism that appears to be a key driver in the development of the vascular system. That mechanism had not previously been described. Yale biologists then validated that discovery in an animal model, proving that our AI method had accurately predicted the role of this particular pathway in vascular development. So, when used correctly, AI can open up a whole new druggable pathway.

Challenges and hurdles
We must look at AI as a force-multiplier, as opposed to a replacement for independent thinking. That’s one reason having domain expertise in genetics and biology is absolutely key. The second thing is that you have to have access to a wealth of information, including cohort studies as well as genomic and phenotypic data.

The role of AI in evolving drug discovery and development?
Much like any major technological development, it’s going to start up slowly and then gather momentum. We believe that because of AI’s unique ability to bring large and complex data sets together and identify patterns within them, that in the end, it’s going to have an exponential impact in advancing and applying precision medicine. The result is going to be game-changing benefits to patients around the globe in terms of better diagnostics and better-targeted drugs.

Read the full interview.