John Garcia

Is better patient care dependent on sharing genetic data?

Invitae scientist John Garcia on reducing VUS rates through data sharing

The inevitable fall of VUS rates

There has been a lot of discussion recently about the rate of Variants of Uncertain Significance (VUSs) in the field of genetic testing. The fact is that some VUSs are unavoidable in genetic testing; every patient has a background rate of genetic variants that cannot be definitively interpreted at this point in our understanding of genetics. However, VUS rates have dropped for all genetic testing companies over time and will continue to drop as research continues and more is learned about the significance of genetic variants.

Thanks to large-scale efforts in the research community such as the 1000 Genomes Project and the NHLBI GO Exome Sequencing Project, scientists have catalogued hundreds of thousands of VUSs that were found in control individuals that are therefore unlikely to cause disease. This allows genetic testing labs to focus on the truly rare VUSs and try to figure out their meaning.

Knowledge is more valuable when shared

As labs sequence more genetic information, the community learns more, but only if this sequencing data is shared with the research and medical communities. For instance, if labs across the world find the same VUS in different patients, those variants will always remain ‘uncertain’ if they remain in a vacuum. However, if several labs independently observe that the VUS does not track with disease in families, then when another lab finds the variant in a patient they can then say that the VUS does not cause disease. And at this point, the VUS is no longer a VUS! This is why it is so important for labs to share their information and not keep their interpretations locked up in their own private silos. Notably, the Global Alliance for Genomics and Health is an international coalition dedicated to sharing clinical genomic data from hundreds of clinical and research labs around the world, and ClinGen is building tools to improve and standardize classifications. By advancing collective knowledge about the impact of different DNA sequence changes, scientists and researchers will improve healthcare for everyone.

Building greater clinical confidence

Data sharing has the added benefit of allowing for independent verification; we can all feel more confident about the meaning we’ve ascribed to a variant if we know that other scientists have observed the same variant and come to the same conclusions. It is also imperative that labs provide the evidence behind their interpretations, because only by understanding why they came to their conclusions can the interpretation be applied to other patients in the future.

What makes a VUS no longer a VUS? Open (and sometimes messy) debate

One natural consequence of all of this sharing of data is that there will always be disagreements, and that’s ok! The state of knowledge is constantly in flux and the interpretations of variants will always change over time. For instance, there is a particular variant in the APC gene known as I1307K (p.Ile1307Lys). It has been suggested that this variant can confer a significantly increased risk of colorectal cancer, but there is a significant amount of controversy about this interpretation. This variant is quite common in the Ashkenazi Jewish population but a very large meta-analysis (which combines the data from lots of smaller studies) showed that Ashkenazi Jewish individuals who have this variant are twice as likely to develop colorectal cancer PMID 23576677. However, the same study also conversely found that individuals who were not of Ashkenazi Jewish ancestry were not at increased risk even if they had this variant. At Invitae, the fact that the association does not hold in both the Ashkenazi and non-Ashkenazi populations leads us to conclude that this sequence change itself does not cause disease. However, we recognize that other labs might interpret things differently and that is why we are sharing our data on this variant to further the discussion. For what it’s worth, none of the individuals who carried the I1307K variant at Invitae had any history of colorectal cancer, including those who reported being of Ashkenazi Jewish decent.

How is Invitae sharing data?

Here at Invitae, it is our mission to further the cause of genetic data sharing. To that end, we fully support the efforts of the ClinVar project to centralize all of the world’s known relationships between genetic variants and disease. ClinVar is a freely available archive of submissions from across the world of variants and the interpretations by the lab who observed them. We have begun to submit our clinically-observed genetic variants to ClinVar, starting with data from a prospective study of high cancer-risk patients in collaboration with Massachusetts General Hospital. Over the next year, we will routinely share as many variants as we can with ClinVar. This is merely the first step on our continuing commitment to sharing our data.

Designing a “Google for Genetics”

As much as we would all like a central repository of all clinically observed variants, some labs have kindly developed their own public variant databases that have not been integrated into ClinVar. To help the community, we have built ClinVitae, a freely available database of clinically observed variants aggregated from ClinVar as well as several other publicly available databases. Within ClinVitae, you can search across all of the databases simultaneously with a single consistent interface. We have also tried to make ClinVitae fast and intuitive enough to be used just like a regular search engine; a “Google for Genetics” if you will.

There is a great deal of information in ClinVar and other databases that we can all use right now to understand genetic variation and impact healthcare. ClinVitae, which is currently in beta, will be updated in the next few months with some new features that will make the cross-database search functionality even easier. We built ClinVitae with an eye towards helping clinicians and genetic counselors quickly find the most relevant information about any variant found in the included databases, and we would like to people to use it and provide feedback.

Get involved in the open-access revolution

Invitae also supports grassroots efforts to publish clinically observed variants from labs who do not routinely share their own data. For example, the BRCA1 and BRCA2 genes potentially represent the most repeatedly sequenced genes in the entire field of genetic testing, and yet the clinical understanding of variation in these genes is impeded by the lack of freely available information.

If you are a patient who has received a lab report with BRCA1 or BRCA2 results, please visit Free the Data to share your report and join a community dedicated to sharing your story with others whose lives have been impacted by breast cancer.

If you are a clinic who has treated people with hereditary breast and ovarian cancer, please visit Sharing Clinical Reports Project where you can easily submit de-identified reports.

Both resources will share the variants with ClinVar, where clinicians and patients around the world can benefit from the information you are providing.

Realizing the full potential of genetic information

The discussion about VUSs is an important one at this point in time, but thanks to crescendoing data sharing efforts, we feel that the next few years will herald a wondrous increase in the medical community’s understanding of the impact of individual genetic changes and their effects on human health. Especially with the advent of next-generation sequencing technologies that will dramatically increased ubiquitousness of genetic testing, the influx of data has never been faster. By pooling our resources, joining forces, and sharing our combined expertise, the medical community will ride the “tsunami” of genetic information described by Invitae’s Jill Hagenkord, towards a greater understanding of genetic variation.

 

2 thoughts on “Is better patient care dependent on sharing genetic data?

Comments are closed.