5 ms·

Why would it not be possible? Considering how quickly the price has been falling [0], I would assume the price to be fairly close to $200-$300 by this point. So

Why would it not be possible? Considering how quickly the price has been falling [0], I would assume the price to be fairly close to $200-$300 by this point. Some are even targeting $100 soon [1]. It's also worth noting that this is an extreme sale, so they may be running at a loss or no profit, mostly for brand building. I definitely hadn't heard of them before today. [0] https://www.genome.gov/27541954/dna-sequencing-costs-data/ https://www.genome.gov/27541954/dna-sequencing-costs-data/ [1] https://techcrunch.com/2017/01/10/illumina-wants-to-sequence-your-whole-genome-for-100/ https://techcrunch.com/2017/01/10/illumina-wants-to-sequence...

rfc 8y ago
Yeah... it unfortunately doesn't work like that. Illumina is notorious for glossing over many of the finer points around how they get to the fictitious numbers they claim. What they don't point out are all of the reagents, prep kits, overhead, lab costs, labor costs, software costs, pipettes, etc. The only way this company stays alive without significant venture funding (which it appears they don't have) is through partnering with pharmas. From reading their privacy policy, it looks like they are de-identifying genotypic and phenotypic data from users. This, at scale, creates a nice population based genetics study so that if I'm a pharma, I can look at a specific biomarker and all data associated with it. That helps me build drugs against certain genes, diseases, etc. So yes, you are the still the product and yes, your data is being resold.
- jghn 8y ago
  They’re using BGI machines, not Illumina
  - rfc 8y ago
    They don't have their own labs based on how they phrased it within their About Us section. Given that + the turnaround time I'm seeing on the web, I'm willing to bet they outsource this to universities who have sequencer downtown. Also, in their About Us section they have every technology listed under the sun (Illumina, PacBio, Thermo, BGI) so I think it's just whatever they can get their hands on based on which labs have open runs. Seems quite questionable from a data integrity perspective if you're trying to do population based genetics. The quality between sequencers varies dramatically.
    - new299 8y ago
      Given the error rates shown for this product it seems unlikely this is anything but BGI sequencing. The BGI (MGI) have been pushing out new instruments, reportedly with lower error rates (and much cheaper than Illumina). They’ve also been building out service centers outside China, and I suspect this is where the sequencing being done by this service is located. Universities do sometimes offer sequencing as a service, but I suspect not at as low a price point as the BGI, even for those universities with attached genome centers.
      - rfc 8y ago
        Good catch on the error rates. Didn't see that. We go up against universities all the time in our competitive deals. They're generally always cheaper since everything is subsidized however the turnaround times are usually significant and the quality is not entirely repeatable. The cost here looks heavily subsidized just as a way to spark interest.
- soared 8y ago
  If data is anonymized properly, aren't we just enabling the creation of better, more effective medicine?
  - oarsinsync 8y ago
    Define "properly". At one point, we thought that simply replacing a user's IP with a unique ID was enough to anonymise. At one point, a blur or pixelating filter on a photo was sufficient to anonymise it. At one point, Monero transactions were considered anonymous.
  - cure 8y ago
    There is nothing that identifies a person more precisely than their genome. It is, after all, unique to each person. That means that by definition, anonymizing genomes is impossible. If someone tells you they will 'anonymize' your genome data, run away, they don't know what they are doing.
  - jacquesm 8y ago
    You can't anonymize your DNA.
- new299 8y ago
  Illumina’s margins on consumables are very high: http://41j.com/blog/2018/11/illumina-consumables-are-90-profit/ http://41j.com/blog/2018/11/illumina-consumables-are-90-prof... Their overall gross margin is >60%. I suspect that instrument margins are actually also quite high, but they factor in R&D spending on instrument costs. However, Dante labs appear to be using BGI, not Illumina for sequencing.
klmr 8y ago
Real numbers for Illumina sequencing currently hover at around 600 USD per 30x coverage human WGS (and this is pure sequencing cost, excluding data storage and analysis). Prices are not dropping since Illumina has no real incentive at the moment. That said, Dante Labs is using a competitor (BGI) who claims to be doing cheaper sequencing (at reduced quality). Even so, Dante is currently definitely operating at a loss. That’s not a problem though since their business model probably isn’t really consumer genetics. Likely, their actual business model (like that of every major player in the field) is gathering large panels of sequence data to either analyse themselves (and then sell the insight to pharmaceutical companies) or to sell (pseudonymised or aggregate) access to companies. So, in reality, they are preparing large-scale biomedical studies. And not only are they not paying study participants (their ostensible “customers”), they are actually getting the study participants to pay for their participation. (Note that there’s nothing necessarily nefarious about this whole business. It just explains why they can offer their product to consumers “at a loss”. In reality it’s simply an investment.)

rfc 8y ago

Yeah... it unfortunately doesn't work like that. Illumina is notorious for glossing over many of the finer points around how they get to the fictitious numbers they claim. What they don't point out are all of the reagents, prep kits, overhead, lab costs, labor costs, software costs, pipettes, etc. The only way this company stays alive without significant venture funding (which it appears they don't have) is through partnering with pharmas. From reading their privacy policy, it looks like they are de-identifying genotypic and phenotypic data from users. This, at scale, creates a nice population based genetics study so that if I'm a pharma, I can look at a specific biomarker and all data associated with it. That helps me build drugs against certain genes, diseases, etc. So yes, you are the still the product and yes, your data is being resold.

jghn 8y ago
They’re using BGI machines, not Illumina
- rfc 8y ago
  They don't have their own labs based on how they phrased it within their About Us section. Given that + the turnaround time I'm seeing on the web, I'm willing to bet they outsource this to universities who have sequencer downtown. Also, in their About Us section they have every technology listed under the sun (Illumina, PacBio, Thermo, BGI) so I think it's just whatever they can get their hands on based on which labs have open runs. Seems quite questionable from a data integrity perspective if you're trying to do population based genetics. The quality between sequencers varies dramatically.
  - new299 8y ago
    Given the error rates shown for this product it seems unlikely this is anything but BGI sequencing. The BGI (MGI) have been pushing out new instruments, reportedly with lower error rates (and much cheaper than Illumina). They’ve also been building out service centers outside China, and I suspect this is where the sequencing being done by this service is located. Universities do sometimes offer sequencing as a service, but I suspect not at as low a price point as the BGI, even for those universities with attached genome centers.
    - rfc 8y ago
      Good catch on the error rates. Didn't see that. We go up against universities all the time in our competitive deals. They're generally always cheaper since everything is subsidized however the turnaround times are usually significant and the quality is not entirely repeatable. The cost here looks heavily subsidized just as a way to spark interest.
soared 8y ago
If data is anonymized properly, aren't we just enabling the creation of better, more effective medicine?
- oarsinsync 8y ago
  Define "properly". At one point, we thought that simply replacing a user's IP with a unique ID was enough to anonymise. At one point, a blur or pixelating filter on a photo was sufficient to anonymise it. At one point, Monero transactions were considered anonymous.
- cure 8y ago
  There is nothing that identifies a person more precisely than their genome. It is, after all, unique to each person. That means that by definition, anonymizing genomes is impossible. If someone tells you they will 'anonymize' your genome data, run away, they don't know what they are doing.
- jacquesm 8y ago
  You can't anonymize your DNA.
new299 8y ago
Illumina’s margins on consumables are very high: http://41j.com/blog/2018/11/illumina-consumables-are-90-profit/ http://41j.com/blog/2018/11/illumina-consumables-are-90-prof... Their overall gross margin is >60%. I suspect that instrument margins are actually also quite high, but they factor in R&D spending on instrument costs. However, Dante labs appear to be using BGI, not Illumina for sequencing.

klmr 8y ago

Real numbers for Illumina sequencing currently hover at around 600 USD per 30x coverage human WGS (and this is pure sequencing cost, excluding data storage and analysis). Prices are not dropping since Illumina has no real incentive at the moment. That said, Dante Labs is using a competitor (BGI) who claims to be doing cheaper sequencing (at reduced quality). Even so, Dante is currently definitely operating at a loss. That’s not a problem though since their business model probably isn’t really consumer genetics. Likely, their actual business model (like that of every major player in the field) is gathering large panels of sequence data to either analyse themselves (and then sell the insight to pharmaceutical companies) or to sell (pseudonymised or aggregate) access to companies. So, in reality, they are preparing large-scale biomedical studies. And not only are they not paying study participants (their ostensible “customers”), they are actually getting the study participants to pay for their participation. (Note that there’s nothing necessarily nefarious about this whole business. It just explains why they can offer their product to consumers “at a loss”. In reality it’s simply an investment.)