Hi Alan,
Looks interesting as I have my data from 23andme, but also a lot of questions here, some of the most obvious are:
1) Credibility. As health information is VERY sensitive, then, first of all, I will try to evaluate the credibility and accuracy of the recommendations you provide. From how your algorithms work to whether your team or organization has the required expertise and training to provide that kind of information, and more!
2) Why would I need a monthly subscription?
At this point from the home page and "how it works" that was not very clear to me how it ACTUALLY works. The explanatory video is quite long to get the first quick understanding of the product (but could be good when I'm ready to learn all the details about it).
Hope to find the answers and actually try it out!
Good luck!
Report
@aleks_muse
Hi Aleksandra -
1) we curate published information that is available on both public and paid databases, and then our team of experts (with graduate degrees in life sciences) manually edit and curate this information to ensure correctness and accessibility. In the near future, we are looking at using more advanced data analysis tools including machine learning and data mining.
2) Because features are being released on a regular basis. In addition, new studies are being published every day (which we don't have control of, but we do update our database regularly) so what you get today could be different from what you get tomorrow.
@datarade you are right, and we are aware of that. Just that the SNPs are the most accessible information right now. This is why we don't stick to a few SNPs with most studies on them, but we rather analyze all the data instead. We do tell people that SNPs isn't everything because there's a lot more to genetic expression and phenotypes, but having this data and access to the analysis is empowering. This is why we have pathway tools to look at all the genes together in context with the symptoms.
Maybe someone could criticize 23andme for not releasing the real raw data that would make CNVs available... lol those could actually be a lot more informative than the SNPs. If you are familiar with this sort of data, it's possible to extract CNV data from the signals from this chip.
@datarade came into this product just to CTRL+F and find your comment
Report
A lot of claims..
How can you guarantee that the outputs are correct (based on what, exactly) and what kind of guarantees can you offer that these recommendations are going to produce what they claim?
Is it a continuous subscription in exchange for a bet on the recommendations with no commitment to end results?
"Tied to longer descriptions" seems to me like a hook into a more prolonged betting on.. "recommendations".
Can you offer a bit more transparency about your methods, approach and plan?
Report
@lyondhur we curate published data from public (e.g. PubMed) and paid databases, so to be precise, we broker data, put them together in accessible formats, and have a team of experts curating them. It wouldn't be legal to make guaranteed health claims (neither could any health practitioner or other piece of software). This is a tool to help people take a look at their genetics in context and provide a few actionable things to try. In the end, the only way to know if something will work for them is to try it (with some safety precautions in mind, of course). This is the tenets of biohacking and SelfHacked by definition.
Report
@natchamaithai thanks for the prompt response Nattha. I wished I were able to see more drilling into those questions; I do appreciate it though.
As you yourself have voiced, SelfDecode would be a recommendation platform (as is, that's not very clear). It looks as though you're providing a genetic improvement service of advice, instead of what I believe your intention to be: decoding genetic data and specialised help translating and framing it, so that people can do whatever they want with it. You might want to look at they way the website presents the opportunity to the public.
Near all of us in here are fairly capable of identifying what biohacking means, the generalities of your business operation and part in manipulating public/private datasets, as well as the academic weight terms like specialists and PhDs play on reinforcing to the general public that "SelfDecode may improve someone's health" (that's sort of the tone here). The general public needs way more simplified formalisation around this.
Report
@natchamaithai
As mentioned by others, SelfDecode also encourages people to share their symptoms - not only their genetic descriptors from other sources - in some sort of an internal forum. Again, with a TOS of only three small paragraphs where phrasing such as "unintentionally leak/share your info with a third party" its ever a bit hard to understand your entire direction. There's a real privacy problem here that needs answering. I can't see any of it.
No doubt the science is mustard. However, It seems to me you guys are going to the public with quite a few other knots to tie out before that. "Genetic Digital Health Coach" is sooooo ambiguous...
I hope all this may be taken with an opened mind.
Good luck with the project.
Report
@lyondhur It is a recommendation platform and there are many different ways to call it. Maybe we can improve on the communication part? Feel free to throw some ideas at me. I think this is helpful. Sometimes we nerds get caught in our own heads and the nerdier details so we forget the simple stuff like this.
Since DNA data is surely the most personal data you can gather, I wonder how do you ensure data security? Are you HIPAA certified? I would not upload any DNA data to any service like that without detailed information on privacy and security...
Report
@josef_moser we are not HIPAA compliant (as would any other DNA analyzer) but we make reasonable efforts to protect personal information with data encryption and only allowing data access for a few key staff in the company. I didn't feel safe using my own name when I submitted my samples to 23andme because the laws with respect to this type of data are still developing. To point out another risk, your genetic data can give you information about your blood relatives, too. In the end, you just have to weigh the risks and benefits, but I think using a pseudonym is the way to go.
Report
How is better than Promethease? Why would I need a monthly subscription rather than paying a one-time fee?
Report
SelfDecode has much more data than Promethase and has precise, actionable outputs. SNPs are then tied to longer descriptions from sister company SelfHacked so you truly understand what's happening in your genome. SelfDecode is morphing to include symptoms, lab wok, community and more. So we're looking at a unique, person-specific way of hacking your biology in a way that's not really been done before. And I'm not overzealous ;)
Joe Cohen is my primary trusted source of biohacking advice via his site: https://selfhacked.com/
SelfDecode is a super interesting new product, which can give you really interesting insights based on your DNA. It's like 23andme, but with actionable outputs.
Report
@_jacksmith Yep, actionable outputs which are feedback optimized FTW
Report
I'm very much into learning about genes right now so overall its one of my favorite sites and very very useful. Definitely my go-to along with their companion site selfhacked. Its always interesting just looking around.
I like the fact studies are linked to any assertions the sites makes and i can easily corroborate and get a different take, if any, on the info by my own searches. Its also very helpful to look up by disease, mutation, snp or medicine although not always successful.
Overall the site frames the information that otherwise is greatly difficult to find or not available in other snp decoders in a more comprehensive way than other sites. I've used almost all the other sites and they were mostly the same and while very useful in their way, just didn't have the breadth and depth this one does.
Its also very interesting to look at the population studies which gives one an idea about the percentage of population each allele is supposed to have. I am interested in finding out how you do that math on that. I'm also curious how you are approaching building up your database/reasearch? Are you focusing on specific areas?
I have found it very very helpful in understanding which snps have mutations, if any. It gives users a better vocabulary about genes which then makes it easier to read the linked NIH studies or at least know what terms to use if searching for more information. There are still holes in the ordering - aka if some snp is listed as a 'potentially bad gene' it's still high up on your bad genes list even if your own snp alleles don't contain those risk alleles.
thanks for all the work thats been put in. I look forward to your progression.
Product Hunt
Product Hunt
Link Texting
Timber
Toolhouse
Cryptoradar
Hustle X