If anyone is curious about genomes: it’s unlikely. Veritas only just started offering a $1k genome; no one is announcing something that’ll be ready by June which is in 5 months; and if you want to extrapolate from the historical data, while we’re clearly recently jumped to a new regime starting in 2015, but extrapolating from the 2015-2016 data, genomes still shouldn’t be <$500 for another 10 months or so:
(10 rather than 8 because the first 3 time-units, 1-3, have already passed, and the remaining 5 time-units until $471 is in time-units of 2-months, so 2*5=10. I couldn’t be bothered to convert the data to more sensible days/months for a quick extrapolation.)
So I would expect any $500 genome to be in 2017 or 2018. Some dark horse could overnight announce a $500 whole-genome at any time… but I wouldn’t bet on it, in part because that NHS data may not reflect any such service’s existence.
There is a small issue that NIH is quoting a wholesale price, while Veritas is quoting a retail price (including customer acquisition, risk of not using all capacity, and profit), so you should expect the NIH price to be lower.
So I would expect any $500 genome to be in 2017 or 2018. Some dark horse could overnight announce a $500 whole-genome at any time… but I wouldn’t bet on it, in part because that NHS data may not reflect any such service’s existence.
Yes. The metric that’s used is NIH data. Given that they brought sequencing machines in the last years and the cost of the sequencing machine’s isn’t completely included in the year they are brought and the 2016 sequencing budget likely pays for machines brought in 2014 and 2015 even if great new machine get’s introduced in 2016.
Right. Great internal validity since you can count on them to not be playing any games with costs like, say, commercial companies such as Illumina; but imperfect external validity which renders down-to-month precision questionable.
Top rated questions on the new prediction website metaculus:
Fully autonomous self-driving cars by 2018?
Will “Planet Nine” be discovered in 2016?
In 2016, will an AI player beat a professionally ranked human in the ancient game of Go?
Will the cost of sequencing a human genome fall below $500 by mid 2016?
Experimental tests of quantum effects in cognition?
If anyone is curious about genomes: it’s unlikely. Veritas only just started offering a $1k genome; no one is announcing something that’ll be ready by June which is in 5 months; and if you want to extrapolate from the historical data, while we’re clearly recently jumped to a new regime starting in 2015, but extrapolating from the 2015-2016 data, genomes still shouldn’t be <$500 for another 10 months or so:
(10 rather than 8 because the first 3 time-units, 1-3, have already passed, and the remaining 5 time-units until $471 is in time-units of 2-months, so 2*5=10. I couldn’t be bothered to convert the data to more sensible days/months for a quick extrapolation.)
So I would expect any $500 genome to be in 2017 or 2018. Some dark horse could overnight announce a $500 whole-genome at any time… but I wouldn’t bet on it, in part because that NHS data may not reflect any such service’s existence.
There is a small issue that NIH is quoting a wholesale price, while Veritas is quoting a retail price (including customer acquisition, risk of not using all capacity, and profit), so you should expect the NIH price to be lower.
Yes. The metric that’s used is NIH data. Given that they brought sequencing machines in the last years and the cost of the sequencing machine’s isn’t completely included in the year they are brought and the 2016 sequencing budget likely pays for machines brought in 2014 and 2015 even if great new machine get’s introduced in 2016.
Right. Great internal validity since you can count on them to not be playing any games with costs like, say, commercial companies such as Illumina; but imperfect external validity which renders down-to-month precision questionable.