Not many people understand the nature of this work and its possible uses.
Mapping the genome will primarily fuel research, making the life of researchers much easier. What this mapping has achieved is to find the areas of DNA that actually contain information. Thus, many billions of DNA bases have been ruled out as the cause for disease (even if they do play a role in supporting the overall DNA structure etc)
The search for genetical variations proceeds in parallalel at a different level, and already one can try to search for e.g. p53 protein variants in the GenBank. Less important proteins might not show, however, but they and their variants will be discovered in time if the particular genetic locus has medical significance.
It is already possible (but not economically feasible) to monitor a newborn for hundreds of diseases but this is extremely costly and the diseases are so rare, that we would be spending (quoting a calculation from a book) e.g. 350K dollars in prenatal and other checks to find a single child with cystic fibrosis (if I remember correctly). All diseases caused by a single gene, esp. the fatal ones, are so damn rare that there is no point in using existing technology to find them.
On the other hand, common diseases are promoted by a multitude of factors, possibly including genetic variations (susceptibility to smoke causing obstructive lung disease, susceptibility to alcool causing cirrhosis). These diseases are frequent, but the genetic background is not strong and not causative, and therefore it is difficult to find responsible genes. IF, however, we had the ability to detect variations that would make a person more sensitive to smoke then we could tell this particular person to quit smoking and that would be very effective.
I believe that the main use (and propably the only ethical one) of genetic screening would be to discover sensitive subpopulations and allow for efficient preventive medicine in these populations.
(note that the affordability of preventive measures always has to do with the frequency of the disease! if we knew the high-risk subpopulation then it would much more cost effective to apply preventive measures to it, than to the general population, because the frequency of the disease is high in the sensitive group and preventive measures "pay off")
As for the possibility to >curecurrent uses of genetic knowledge in everyday medicine. I'm quite sure that the HG map will be an important step in integrating low level molecular data with high level clinical response and treatment.
This is not necessarily true. Algorithmic improvement is extremely important. The original Fourier Transform (O(N^2)) vs the Fast Fourier Transform (O(N*log(N))) is dramatically slower, even though for small N, simpler code may make the N^2 algorithm complete in less time.
I have seen programs run faster on a 386 than on a Pentium II, because they used better algorithms. No amount of assembly art-work can give you that.
I think that saying "buy a better processor to make slow software run fast" encourages bloated and low quality software. I strongly disagree with that. Software (algorithms) makes the difference.
D. E. Knuth (if I remember correctly) has spoken against patenting algorithms and maybe you should have a look at his page. I agree with him (or whoever it was, I read about that many years ago).
Slashdot Mirror
Not many people understand the nature of this work and its possible uses.
Mapping the genome will primarily fuel research, making the life of researchers much easier.
What this mapping has achieved is to find the
areas of DNA that actually contain information.
Thus, many billions of DNA bases have been ruled out as the cause for disease (even if they do play
a role in supporting the overall DNA structure etc)
The search for genetical variations proceeds in
parallalel at a different level, and already one
can try to search for e.g. p53 protein variants
in the GenBank. Less important proteins might not
show, however, but they and their variants will be discovered in time if the particular genetic locus has medical significance.
It is already possible (but not economically
feasible) to monitor a newborn for hundreds of
diseases but this is extremely costly and the
diseases are so rare, that we would be spending
(quoting a calculation from a book) e.g. 350K
dollars in prenatal and other checks to find a
single child with cystic fibrosis (if I remember
correctly). All diseases caused by a single
gene, esp. the fatal ones, are so damn rare that
there is no point in using existing technology
to find them.
On the other hand, common diseases are promoted by
a multitude of factors, possibly including genetic
variations (susceptibility to smoke causing obstructive lung disease, susceptibility to alcool
causing cirrhosis). These diseases are frequent,
but the genetic background is not strong and not
causative, and therefore it is difficult to find
responsible genes. IF, however, we had the ability
to detect variations that would make a person more
sensitive to smoke then we could tell this particular person to quit smoking and that would be very effective.
I believe that the main use (and propably the only
ethical one) of genetic screening would be to discover sensitive subpopulations and allow for
efficient preventive medicine in these populations.
(note that the affordability of preventive measures always has to do with the frequency of
the disease! if we knew the high-risk subpopulation then it would much more cost effective to apply preventive measures to it, than to the general population, because the frequency of the disease is high in the sensitive group and preventive measures "pay off")
As for the possibility to >curecurrent uses of
genetic knowledge in everyday medicine. I'm quite
sure that the HG map will be an important step
in integrating low level molecular data with
high level clinical response and treatment.
Petros
P.S. My message looks awful. Shit.
This is not necessarily true. Algorithmic improvement is extremely important. The original
Fourier Transform (O(N^2)) vs the Fast Fourier
Transform (O(N*log(N))) is dramatically slower,
even though for small N, simpler code may make
the N^2 algorithm complete in less time.
I have seen programs run faster on a 386 than on
a Pentium II, because they used better algorithms.
No amount of assembly art-work can give you that.
I think that saying "buy a better processor to make slow software run fast" encourages bloated
and low quality software. I strongly disagree with
that. Software (algorithms) makes the difference.
D. E. Knuth (if I remember correctly) has spoken
against patenting algorithms and maybe you should
have a look at his page. I agree with him (or whoever it was, I read about that many years ago).