NEW YORK (GenomeWeb) – With growing recognition in the lab community that the accuracy and safety of genomic analyses may depend on shared access to available data, sharing tools and programs have abounded in recent years.
However, this movement has not translated into direct accessibility for patients and families, despite the fact that recipients of clinical genomic analyses have a legal right to their test results.
To address this and "democratize" genomic data, a University of Washington team earlier this year launched MyGene2, a web tool to facilitate gene discovery, genotype-phenotype understanding, and eventually diagnoses among families with suspected Mendelian disorders.
Mike Bamshad and Jessica Chong, who led development of the platform, spoke with GenomeWeb this week about how MyGene2 has advanced over the past year, and what is next for the tool.
Bamshad runs one of several Centers for Mendelian Genetics (CMGs), where he has seen growth of family-directed, social media-based efforts to link up cases of rare diseases in order to find genetic diagnoses for their children.
A notable early example was the story of Matt Might, who in an attempt to find out if the NGLY1 mutation his son carried was the cause of his symptoms, wrote a blog post that went viral, bringing together some half a dozen other patients with NGLY1 mutations and their families.
There have been other similar success stories, Bamshad said, some of which he and colleagues reported on in Nature last year. "They have told us, 'you guys should develop a tool to make this easier for other families to do,'" he added. "We actually proposed something like that 5 years ago [amongst the CMG consortium,] so we thought maybe it's time to implement it."
While there are databases that researchers can search to make new genotype-phenotype matches, results are rarely accessible to patients themselves and in some cases matches aren't even necessarily public to other researchers. And while groups work to publish their findings, this takes so much time that in high-volume operations like the CMGs new discoveries can go unpublished for months or years, remaining siloed in a single institution and inaccessible to others who might have other cases that fit the same profile.
MyGene2 was one of six finalists in the first round of a competition called the Open Science Prize, standing out from a pool of more than 90 applicants, and receiving an initial award of $80,000 this May.
Along with their five competitors for the final round and grand prize of $230,000, Bamshad and Chong last week presented their vision for MyGene2. Public voting will play into selection of the prize winner, and interested parties can vote here.
Currently, the platform includes a database for sharing health information and genetic data that is searchable not only by researchers, like many current genomic databases, but by any family or individual interested. They can also sign up for notifications of new discoveries, and network with other users.
Individuals input not only their own genomic sequencing or other test results into the system, but also descriptions of their physical symptoms or phenotype. Doctors and researchers can also share de-identified data. All of this information can then be compared and analyzed by researchers, but also by families with rare diseases, if desired.
According to Bamshad, the site now has more than 500 profiles, about half of which were created by families or individuals and another half by researchers. The database includes more than 470 candidate genes.
When MyGene 2 became a finalist for the Open Science Prize earlier this year, its developers pledged to join the MatchMaker Exchange (MME), a network of restricted access databases designed to share health and genetic data among doctors and researchers to promote discovery.
Presenting for the final round of the competition, the researchers reported that they've met this goal, and all novel candidate genes shared by doctors and researchers in MyGene2 are now automatically shared with MME.
Families have traditionally been forbidden from participating in MME, but the MyGene2 team said it in its report that it has successfully lobbied MME to allow data submission from families, and a system for allowing this is in the works.
The group reported that it is also working on incorporating a tool called gene.iobio, developed by researchers at the University of Utah, to allow families to easily browse their own sequence data.
Finally, the MyGene2 team has also built out an automated pipeline for identifying candidate gene matches, as well as a system to report these matches in real time. The reports summarize health problems found in each family with a mutation in the gene in question.
When a user clicks through to a match report, it lists inheritance models under which variants were identified, who submitted each case — whether a family or a researcher or clinician — and a potentially long list of physical features, marked as to how many of the reported cases they occur in.
"For all of those genes where two or more families have mutations … we create a match report," Bamshad said. While some of these might reflect known disorders or associations, others are novel.
For example, the site's gene list includes a match report for the gene SRPK3. A "Y" in the "novel gene" column indicates that the match is a new one, i.e. it hasn't been reported elsewhere in the scientific literature.
Another gene on this list is MYLP, which the UW team entered using data from its own sequencing efforts that picked up two cases with the same variant. "We put that in to share with other researchers, because we are very confident that it's a real association," Bamshad said.
As the group has progressed toward the second round of the Open Science Prize challenge, other improvements have included a feature whereby families can assign a doctor or genetic counselor as a proxy to assist with the entry of their health and genetic information, provide guidance about matches, and moderate communication with researchers.
Through the MyGene2 website, families who don't yet have genomic data can also now apply directly for research exome sequencing by the UW Center for Mendelian Genomics or for self-paid CLIA-certified research exome sequencing.
Though not all research sequencing groups make data available to participants, this is also a growing trend. At UW, Bamshad and colleagues have developed another web-based tool, called My46, which allows individuals who are sequenced in a research context to get access to their results through a web portal.
Between MyGene2 and My46, families can seek out exome sequencing, get results, and have that raw data imported back for sharing with others.
Moving forward, Bamshad said the main goals for MyGene2 center around growing participation.
The team has some plans for integrating the platform into both academic and industry workflows. Some of these are more general. For example, at Seattle Children's Hospital, where Bamshad is on staff, patients now get a handout about the importance of data sharing.
More directly, the group also hopes to partner with genetic testing laboratories to advertise MyGene2 to their users and to develop direct data submission from each lab.
Bamshad said he couldn't detail specific potential industry players, but said that companies the group has spoken to so far "see the merit."
Although many have embraced the power of more open data in accelerating discovery and improving interpretation of genetic test results, there are also groups and industry players that have taken the opposite position, most notably Myriad Genetics.
While MyGene2 might not be able to persuade some companies to directly encourage patients to share their data, Bamshad reiterated that anyone who has been tested clinically has legal access to their data, and could request it and share it if they so desire.
The group could also promote MyGene2 by working with advocacy groups for specific diseases or disease areas, or by developing ways that peer-reviewed papers or discoveries in the context of large-scale gene discovery programs can be automatically or directly submitted into the database. These are all things the team is working on.
"The model is highly dependent on users sharing their data," Bamshad said. "Without a really compelling reason for people not to share their data, we feel like if we had a labor force and huge marketing plan we could imagine hundreds of thousands of people creating profiles."
Realistically, though, he said, the team hopes it can push for reaching 10,000 profiles in the next year.