Approval for Dash Genomics

I’m posting a review request for Dash Genomics, which is applying for project approval on Open Humans.

Overview
Dash Genomics has partnered with UCSF and UCSD scientists to offer a new model to understand individual risk of Alzheimer’s disease beyond only the APOE gene, as a function of age and genetic information. Dash Genomics has now made this state-of-the-art algorithm available to anyone with 23andMe or Ancestry.com DNA data.

The personalized Alzheimer’s risk analysis is available from Dash Genomics for $149 but is currently available to Open Humans members at a special price of $99.

Overview of Risk Analysis Offering

  1. Please see www.dashgenomics.com for more information and an interactive sample report.

  2. Report example: https://www.dashgenomics.com/#demoreport

  3. Product page for full description: https://www.dashgenomics.com/#product

  4. We offer the following aspects of analysis within the report

  5. Cumulative Risk – the risk of onset of Alzheimer’s by a certain age

  6. Instantaneous Risk – the risk of onset of Alzheimer’s within a certain year of age

  7. Risk over the next 5 years (for those over 55)

  8. AZ50 age – the age your cumulative risk rises above 50%

  9. Polygenic Hazard Score (PHS) percentile – Compare your estimated risk profile to the general population

  10. APOE status – Status based on the genotype combination of the two APOE copies

How does it work?

Participants can purchase the risk analysis through the Dash Genomics website. Participants have the option to upload their 23andMe or AncestryDNA data files. Dash Genomics validates the file for compatibility before running the analysis. If the files meet the requirements, participants can proceed to payment, after which the personalized report is generated in seconds. Once the report is created, participants are able to log in at www.dashgenomics.com at any time using their laptop or mobile device to access their interactive, digital report and can also print or download a PDF version.

Data collection details:

  1. Raw uploaded files are stored encrypted
  2. All account information and report results are stored in an encrypted database
  3. User passwords are never stored in clear text. A random hash is generated for each password before storage and must be changed if forgotten or lost
  4. Raw files and generated result data is stored separately from the account information without the names of users
  5. Data can be deleted upon request. We remove the user account, uploaded file(s) associated with the account, and the generated report results

The Science

The proprietary algorithm used by Dash Genomics was developed by HealthLytix, a precision health company in San Diego, and is based on cutting-edge research from UC San Diego and UC San Francisco. The research looks at age-specific risk of developing Alzheimer’s disease based on genetic markers that have never been used before for estimating risk. The model looks at the similarities between genetic profile and the average risk in the group analyzed in the research.

Read more about:

  1. Our research team featured in Wired

  2. The science based on a published paper

My concern, given the commercial nature of the application, is that the data being presented is based on well-proven science and provides real value not available from tools such as Genevieve or Protease.

Well… I can confirm that Genevieve doesn’t attempt to do any polygenic risk analysis. :slight_smile: It doesn’t do risk analysis at all, really, my intended focus was more on mendelian predictions and the strength of evidence for variants reported to cause mendelian disease.

I’m not expert in polygenic risk, but according to the project, this tool is based on a published polygenic risk approach (Chouraki et al, 2016). Furthermore, it’s mentioned in a review (Chasioti et al, 2019), and (just a quick scan) it didn’t seem to be criticized there.

So … as far as I can tell, this isn’t obviously “snake oil”. It’s implementing the results of published research on how to perform polygenic analyses for this particular condition.

Notably, it is not attempting to make an ad hoc polygenic guess from research data for individual variants. This would be a lot more sketchy, without some published methodology about how to combine variant information & testing how well the combined algorithm works. So that’s good!

Great discussion, and thanks for the question! As Mad noted, our service is based on peer-reviewed research in leading scientific journals. In addition to the ones listed, our Alzheimer’s polygenic risk score methodology has been published in PLoS Medicine (Desikan et al, 2017) and Brain. Also, we work directly with these original authors of the paper to provide the service. We’re happy to provide more information on the research as helpful!

Best,
The Dash Genomics Team

One more note - we wanted to share this powerful piece by one of our partners Dr. Rahul Desikan on personalized polygenic scores, and how they may help individuals understand their risk for dozens of conditions, including Alzheimer’s.

This project approval failed to link the activity page in Open Humans, and it’s important to review this from the viewpoint of what a member would experience.

Here is the starting point for a member: https://www.openhumans.org/activity/dash-genomics/

Currently the terms/consent page is really really brief – it just links out to the Dash Genomics site.

Can this instead display a copy of all the things @DashGenomics has written above? (Which is great information! Let’s reuse it? :wink: ) Then, at the end it can say what it currently does (linking out to the terms). I feel like this would be a much better “informed user” experience.

Beyond this, I’m happy to see this shared and would support approve.

1 Like

I inherently don’t have a problem with fee based or commercial activity as long as the relationships are defined and based on good science, the latter has been addressed but not the former.

What I don’t see in the https://www.openhumans.org/activity/dash-genomics/ description is a defined financial relationship or lack thereof with OpenHumans. Just given money is involved it is good to provide a clear defined relationship.

If a part of the discounted fee includes a donation to OpenHumans to sustain the infrastructure and community, then that would be cool too, just make it known.

P.S. Even though there is a redirect, I’d rather see https links used.

1 Like

Thanks for your comment!

There is no financial relationship between Dash Genomics and Open Humans. We wanted to provide a discount from our normal price for the community since we know the group is historically interested in such information, and hopefully a discount could make it more accessible to the community. We have historically provided community partners a discount of this size to help the members. We can add the clarification to our consent language as well, so its known that there is no financial relationship/benefit involved here.

Re: URL, great catch! We will update to https.

@wolfgang8741 – we made an edit to our consent language to add the clarfication that there is no financial relationship. Let us know if that makes sense, or if you have any other recommendations on how we can make it clear!

Thanks!

I think a consensus has been reached that this project is OK to be deployed in Open Humans – i.e. approved.

Following board meeting, we support engaging projects in a way that reflects allowing their participation in an open ecosystem. Allowing the project to be operate in the site, following community feedback and consensus, meets that goal.

However, the organization is against specifically promoting a commercial product, regardless of discount to membership, in the absence of more formal and substantial funding/collaboration with a commercial entity (i.e. with oversight of community/board). So, we won’t be sharing this project via newsletter and blog.