3 ms·

How far away we are from a 100$ genome? I've heard that the cost/performance development rate is 'super-Moorean'.

atg_gcg_aca_aca 9y ago
Here's the Sequencing & Moore's Law plot that I'm tired of seeing in every genome sequencing presentation ever. https://wi-images.condecdn.net/image/W5YlBgPjX36/crop/1620 https://wi-images.condecdn.net/image/W5YlBgPjX36/crop/1620 Since 2013, Illumina's brought the price of a genome to about $1k with the HiSeq X, which essentially scales up their existing tech. No idea what the next few years hold, but I'm excited to see insurance companies beginning to cover genome sequencing as an affordable diagnostic tool.
- tintor 9y ago
  Is there an updated plot going up to 2017?
  - martinpw 9y ago
    https://www.nextbigfuture.com/2017/11/update-on-genome-sequencing-costs.html https://www.nextbigfuture.com/2017/11/update-on-genome-seque... Interestingly the cost had not dropped in the past two years. Also the graph references genome.gov but that site appears to be unavailable at the moment.
- bayesian_horse 9y ago
  The question is also what exactly you get. Illumina's technology gives you very short pieces of DNA. I'd call it confetti sequencing. Pacbio and Nanopore both go for longer sequences. This is important when trying to assemble the pieces to a bigger picture.
tjridesbikes 9y ago
I think I could help answer this. The cost of computation has gone down so much that its just a fraction of the price for your whole genome sequencing. The cost comes from preparation, reagents, and storage. Illumina sells their own reagents, so the cost is really dependent on them. We've worked really hard to get the computation cost of genome sequencing as low as possible, and we just passed a major milestone last year. Source: Software Engineer for a well-known non-profit sequencing lab/center.
michaelbarton 9y ago
I think we're already there. You can probably get around a $50 dollar bacterial genome by using robots and doing 384-well plates of samples at a time to get economies of scale. A lot of the costs are outside of the sequencing itself. You have to extract the DNA from the sample and prepare that DNA for the sequencing platform you're using. These costs are both the reagents needed to do this, and also the associated labour costs, even when done at scale. Perhaps you meant $100 human genome though? However I think roughly the same principles apply.

Here's the Sequencing & Moore's Law plot that I'm tired of seeing in every genome sequencing presentation ever. https://wi-images.condecdn.net/image/W5YlBgPjX36/crop/1620 https://wi-images.condecdn.net/image/W5YlBgPjX36/crop/1620 Since 2013, Illumina's brought the price of a genome to about $1k with the HiSeq X, which essentially scales up their existing tech. No idea what the next few years hold, but I'm excited to see insurance companies beginning to cover genome sequencing as an affordable diagnostic tool.

tintor 9y ago
Is there an updated plot going up to 2017?
- martinpw 9y ago
  https://www.nextbigfuture.com/2017/11/update-on-genome-sequencing-costs.html https://www.nextbigfuture.com/2017/11/update-on-genome-seque... Interestingly the cost had not dropped in the past two years. Also the graph references genome.gov but that site appears to be unavailable at the moment.
bayesian_horse 9y ago
The question is also what exactly you get. Illumina's technology gives you very short pieces of DNA. I'd call it confetti sequencing. Pacbio and Nanopore both go for longer sequences. This is important when trying to assemble the pieces to a bigger picture.

tjridesbikes 9y ago

I think I could help answer this. The cost of computation has gone down so much that its just a fraction of the price for your whole genome sequencing. The cost comes from preparation, reagents, and storage. Illumina sells their own reagents, so the cost is really dependent on them. We've worked really hard to get the computation cost of genome sequencing as low as possible, and we just passed a major milestone last year. Source: Software Engineer for a well-known non-profit sequencing lab/center.

michaelbarton 9y ago

I think we're already there. You can probably get around a $50 dollar bacterial genome by using robots and doing 384-well plates of samples at a time to get economies of scale. A lot of the costs are outside of the sequencing itself. You have to extract the DNA from the sample and prepare that DNA for the sequencing platform you're using. These costs are both the reagents needed to do this, and also the associated labour costs, even when done at scale. Perhaps you meant $100 human genome though? However I think roughly the same principles apply.