DNA sequencing for the masses New fathers always think their newborn baby daughters are beautiful, and Hugh Reinhoff was no exception. Sure, she had giant feet, a cleft uvula, and contractures in some of her fingers. But those thoughts, along with a suspicion that her eyes were oddly widely spaced and her chin was a bit recessive, were easily forgotten in …
The scary truth is that this reflects the history of fingerprinting, from a crime detection point of view. The methods used in fingerprint identification today don't stand up to comparisons across millions of subjects but they still hold a great deal of public confidence. Planting a trace of a third party's dna at a crime scene may not be simple but it is by no means impossible.
This is just the first worm in a very large can.
I just took part in an interview to complete a report for Sky News about a company called 23andMe.com - they are backed by a few VC's and Google are in on it too.
They come at this from a slightly different angle (mainly interested in geographical and migratory genetic history) but produce the same detailed genetic info which can be used for any purpose.
The pack (you spit in a tube and send it off!) costs $999 - remarkable!
It'll be less than $500 in 6 months i bet. of course, i got mine free for volunteering (natch)
Will be interested to find out where my ancestors stomped about!
(alien as i am pretty convinced there is something funny in my family tree!)
I just took part in an interview to complete a report for Sky News about a company called 23andMe.com - they are backed by a few VC's and Google are in on it too.
They come at this from a slightly different angle (mainly interested in geographical and migratory genetic history) but produce the same detailed genetic info which can be used for any purpose.
The pack (you spit in a tube and send it off!) costs $999 - remarkable!
It'll be less than $500 in 6 months i bet. of course, i got mine free for volunteering (natch)
Will be interested to find out where my ancestors stomped about!
(alien as i am pretty convinced there is something funny in my family tree!)
Can somebody please explain how Moore's law is related to DNA sequencing? I'm not a geneticist here, but last I checked, DNA did not have transistors. And since Moore's law states that the number of transistors used to create a processor roughly doubles every 18 months, I fail to see how that is even remotely related to DNA sequencing. Even the alternately-accepted meaning (processor speed roughly doubles every 18 months) isn't even remotely related to DNA sequencing.
For two reasons. To understand why you have to understand how whole genome shotgun sequencing works (the Venter approach). You basically take a genome and blast it into loads of little pieces of random length. The smaller the pieces, the easier they are to sequence. However, you've got to work out where all the sequences go in the genome. If you've only got one copy of the genome, it's impossible, so you blast, say, five copies of the genome, sequence them, and because each copy is split in different places, you end up with areas of overlap which you can use to figure out which piece goes where.
Moore's law is relevant, because although a lot of little pieces are (relatively) easy to sequence, it required a sodding enormous computer to calculate where all the overlaps go. Thus, Moore's law is directly relevant, because the more powerful the computer, the quicker the sequencing can be done, and the fewer original genome copies are required in the first place (because the fewer copies are available, the more computing power is required).
The other reason relates to the sequencing machines themselves. I'm a bit more hazy on the technology here, but basically, something like Moore's law is relevant, because the more dna samples you can fit onto a sample carrier (one technique involves placing them in a square pattern), the more sequences you can do in one go.
"It's certainly a lot easier to plant DNA than fingerprints." ...By Kevin Whitefoot Posted Friday 7th March 2008 13:18 GMT
Kevin,
And a lot easier to disprove if planted in the wrong place. In the right place, it will grow and in the wrong place it will prove itself false.
And who would want to plant DNA anyway? For what Honest to Goodness Reason?
Genome sequencing is very heavily dependent on computing power because after the actual sequencing the data needs to be assembled.
It's like shaking all the words from Shakespeare's complete works into a big pile on the floor and trying to put them back the correct order, but with many, many more words.
Previously, a sequencing run would output tens of megabases (millions of letters) per run. We now have upto a gigabase (a billion letters) *per run* of information that needs assembling and considering a run takes about a day, over a working week that's a lot of data. We've been rather spoilt with processing power lately, but this new technology really pushes our hardware!
It's an analogy, and if anything, it understates the rapid advances in sequencing technology.
In the early years of the International Human Genome Project, DNA sequencing was done using gel electrophoresis, a technique that was heavily labour-intensive and rather prone to errors.
Then in 1998, a new technique called capillary sequencing became available through a machine called the Applied Biosystems Inc. 3700, which allowed greater quantities of DNA to be sequenced at lower cost, with far less human involvement, and with a lower error rate. This ushered in "high throughput" sequencing and allowed the human genome to be sequenced far faster than had originally been thought possible when the Human Genome Project was started.
Sequencing technology has continued to develop extremely rapidly. In the past three or four years, several companies, including 454 Life Sciences, Illumina and ABI, have developed machines that can sequence amounts of DNA equivalent to several entire human genomes in a day. It's reached the point where the real problem is how to store the vast amounts of raw data which these machines produce. The Grauniad ran a good article last week:
http://www.guardian.co.uk/technology/2008/feb/28/research.computing
I'm an astronomer by training, but for the past ten years I've worked at one of the world's leading DNA sequencing centres. If telescopes had developed as fast as DNA sequencing technology, Galileo would have been using 10-metre telescopes with adaptive optics back in the early 17th century.
Many already explained, so just a little storage detail here... I'd like to point out that a few hours of activity from our smallest new generation sequencer I work with here (a 454 FLX) can generate about 50 GB of raw data (images, mostly), depending on the size of the job. The other new-gen sequencer (Solexa) generates 1 TB in a day, I've heard. So even more than CPU, our problem now is storage and back up. Where did I put that USB key again?
Oh, and it all runs on Linux. (at least the 454; most of the older ones used to run on Macs)
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
