you're missing a big point
(5) when doing comparitive genomic analysis it's very very helpful to have organisms that are very differen yet conatain the same genes. It really make the controller regions for a gene stick out like sore thumbs as they remain conserved across genomes yet don't show up on any gene prediction programs. This is a very nice shortcut to understanding how to turn genes on and off.
check out more on fugu sequencing at our website doe fugu sequencing
two things come to mind, one JAVA with faster chips it will become practical to have platform independant software that can actually run at a decent speed. Along the same lines faster chips will enable os emulations to run at a much nicer clip.
the other is "The Future" It may sound stupid but really, the argument that chips are fast enough because "everything I need to run, I have the power to run" has been put forth since the dawn of the computer age. What you really need the power for are those things you need to run that don't exist yet because there's not been powerful enough chips about for people to seriously consider writing them.
mapping also now refers to positioning the segments of dna that were sequenced relative to one another. to sequence something large like a chromosome you need to break it up into ~200kb fragments first in order to have a stable starting material. Next you have to break that 200kb fragment into 3kb and 10kb pieces at some fold coverage (determined by how accurate you wan't to be) which are what is actually sequenced the mapping I do is utterly unrelated to genes and solely consists of figuring out how all those fragments originally fit together but it's still called mapping.
they're also using specialized hardware from paracell I believe. the alpha es40s rock, compaq let me try one yesterday and it rocked, 3x speed of latest sun chips for blast and 4x for for threaded fasta searches.
The HGP will finish "draft" status by next june, my part will be at "draft" very very soon. we're spending tons of time now to figure out if we really are at draft status (3 fold coverage) which is something celera doesn't have to worry about, they can just crank up the pr machine in response to their stock dropping so abysmally without having to worry about peer review as they won't be releasing their data for quite some time.
yeah but there's a difference that the.marketing dweebs aren't just moving in, they're displacing those who were here by driving rents through the roof, doubling in the past two years and san francisco wasn't cheap to begin with. Every week it seems theres another story about some little non-profit going under or moving out of the city (if they can) as they can't afford their rent doubling. This week it's cartoon art museum (which was I think the only museum dedicated to cartooning) next week it's be somebody else. Here is a nice story from sf on the dot.whatevers impact in SOMA.
Imagine Sun switching to PowerPC -- the hardware and software rewrites would be enormous!
And yet sun did just that...announcement was here The trouble was they did it for the PReP powerpc reference platform which was suplanted by the CHRP platform which was delayed forever and sun hasn't ported it to anything else powerpcish (that i'm aware of ).
The whole point of the powerpc port was to show they could change processors/architecture w/o overly enormous costs, kind of a big point of unix too eh?
. Actually, this came up in an earlier/. discussion where it was claimed that the HGP uses a single individual as well.
the HGP uses 4 BAC libraries with some P1s and pacs thrown in, for a total of 6 caltech libraries b,c and d pieter DeJongs libraries RPCI-11 and RPCI-1 and the dupont p1 library(I forget the link, it's handled by a private company now)
Each library was constructed from a different persons tissue
1) much like all the comments here your posting neglects to mention berkeleys fruitfly genome sequencing project that did a vast amount of work and without which celeras data wouldn't have been nearly so useful. it certainly wouldn't have made it to finished so quickly without the mapped BACs would it? which leads to point
2) this crap about celera mapping 90% in one year when the public efforts spent 10 years blah blah blah. This really ticks me off, from the very start the plan with the public effort was to spend the vast majority of time developing the technology and techniques necessary to sequence rapidly and accurately, the accelerated curve has been known for ages and our lab went from sequencing apx 2 mb/year to 20 mb in (I think it was) 98 to over 350 mb now seq stats Considering that massive purchases of 377 sequencers and scientific collaborations by the hgp contributed VASTLY to the development of the 3700 it's rather crass to read the crowing about how celeras kicked ass while the public effort allegedly just sat around twiddling their thumbs. The press releases from the formation of celera at least give credit to the planning of the hgp Since the inception of the Human Genome Project (HGP) in 1990, a major shift in technology has been anticipated that would allow the entire sequence to be completed
3) The HGP is now likely sequencing FASTER than celera, I know the doe has 80 megabases(equivalent to the 3700), sanger has 100 3700s and a ton of 377s, that's only 40% of the genome project and celera has what, 230 3700s? hrmm, rough unsubstantiated calcs would put total human effort near least 450 3700s... ouch! Plus MIT now has more than any other group I believe, (although they aren't all working on human) and there's the vast capacity at washU. while it's true celera has the largest private supercomputer and that will help with assembling, the DOE started the human genome project, is still involved and just happens to have the largest supercomputers period.
4)where you get the 10x oversampling number I haven't a clue. the goals are laid out here and additionally a figure of 6x is generally aimed for before trying to finish the clone, finished is still the bahama definition of I believe no more than 1 error per 10k
yeah, I'll bet you have stock options and I'm sure they don't bias your postings and don't influence your continued use of outdated figures
Slashdot Mirror
errgggg... the better link is
Fugu
you're missing a big point (5) when doing comparitive genomic analysis it's very very helpful to have organisms that are very differen yet conatain the same genes. It really make the controller regions for a gene stick out like sore thumbs as they remain conserved across genomes yet don't show up on any gene prediction programs. This is a very nice shortcut to understanding how to turn genes on and off.
check out more on fugu sequencing at our website doe fugu sequencing
JAVA
with faster chips it will become practical to have platform independant software that can actually run at a decent speed. Along the same lines faster chips will enable os emulations to run at a much nicer clip.
the other is
"The Future"
It may sound stupid but really, the argument that chips are fast enough because "everything I need to run, I have the power to run" has been put forth since the dawn of the computer age. What you really need the power for are those things you need to run that don't exist yet because there's not been powerful enough chips about for people to seriously consider writing them.
me.
mapping also now refers to positioning the segments of dna that were sequenced relative to one another.
to sequence something large like a chromosome you need to break it up into ~200kb fragments first in order to have a stable starting material. Next you have to break that 200kb fragment into 3kb and 10kb pieces at some fold coverage (determined by how accurate you wan't to be) which are what is actually sequenced
the mapping I do is utterly unrelated to genes and solely consists of figuring out how all those fragments originally fit together but it's still called mapping.
they're also using specialized hardware from paracell I believe. the alpha es40s rock, compaq let me try one yesterday and it rocked, 3x speed of latest sun chips for blast and 4x for for threaded fasta searches.
The HGP will finish "draft" status by next june, my part will be at "draft" very very soon. we're spending tons of time now to figure out if we really are at draft status (3 fold coverage) which is something celera doesn't have to worry about, they can just crank up the pr machine in response to their stock dropping so abysmally without having to worry about peer review as they won't be releasing their data for quite some time.
yeah but there's a difference that the .marketing dweebs aren't just moving in, they're displacing those who were here by driving rents through the roof, doubling in the past two years and san francisco wasn't cheap to begin with. Every week it seems theres another story about some little non-profit going under or moving out of the city (if they can) as they can't afford their rent doubling. This week it's cartoon art museum (which was I think the only museum dedicated to cartooning) next week it's be somebody else.
Here is a nice story from sf on the dot.whatevers impact in SOMA.
And yet sun did just that...announcement was here
The trouble was they did it for the PReP powerpc reference platform which was suplanted by the CHRP platform which was delayed forever and sun hasn't ported it to anything else powerpcish (that i'm aware of ).
The whole point of the powerpc port was to show they could change processors/architecture w/o overly enormous costs, kind of a big point of unix too eh?
the HGP uses 4 BAC libraries with some P1s and pacs thrown in, for a total of 6
caltech libraries b,c and d
pieter DeJongs libraries RPCI-11 and RPCI-1
and the dupont p1 library(I forget the link, it's handled by a private company now)
Each library was constructed from a different persons tissue
jor-el
1) much like all the comments here your posting neglects to mention berkeleys fruitfly genome sequencing project that did a vast amount of work and without which celeras data wouldn't have been nearly so useful. it certainly wouldn't have made it to finished so quickly without the mapped BACs would it? which leads to point
2) this crap about celera mapping 90% in one year when the public efforts spent 10 years blah blah blah. This really ticks me off, from the very start the plan with the public effort was to spend the vast majority of time developing the technology and techniques necessary to sequence rapidly and accurately, the accelerated curve has been known for ages and our lab went from sequencing apx 2 mb/year to 20 mb in (I think it was) 98 to over 350 mb now seq stats
Considering that massive purchases of 377 sequencers and scientific collaborations by the hgp contributed VASTLY to the development of the 3700 it's rather crass to read the crowing about how celeras kicked ass while the public effort allegedly just sat around twiddling their thumbs. The press releases from the formation of celera at least give credit to the planning of the hgp
Since the inception of the Human Genome Project (HGP) in 1990, a major shift in technology has been anticipated that would allow the entire sequence to be completed
3) The HGP is now likely sequencing FASTER than celera, I know the doe has 80 megabases(equivalent to the 3700), sanger has 100 3700s and a ton of 377s, that's only 40% of the genome project and celera has what, 230 3700s? hrmm, rough unsubstantiated calcs would put total human effort near least 450 3700s ... ouch! Plus MIT now has more than any other group I believe, (although they aren't all working on human) and there's the vast capacity at washU.
while it's true celera has the largest private supercomputer and that will help with assembling, the DOE started the human genome project, is still involved and just happens to have the largest supercomputers period.
4)where you get the 10x oversampling number I haven't a clue. the goals are laid out here and additionally a figure of 6x is generally aimed for before trying to finish the clone, finished is still the bahama definition of I believe no more than 1 error per 10k
yeah, I'll bet you have stock options and I'm sure they don't bias your postings and don't influence your continued use of outdated figures