Posted by Larry Hoover on January 3, 2004, at 11:49:01
In reply to Linkage and Lar - Any thoughts on this article, posted by BobS. on January 1, 2004, at 19:36:26
Hey, Bob. I hope you don't mind the delay. I wanted to wait a while, and read it again before I answered.
Overall, this was a very balanced summary of what the study showed, and what it didn't. However, it does raise some ideas I'd like to discuss more fully.
> The enzyme has at least 40 genetic variations, but to their surprise, the researchers found that the variation did not alter treatment outcome or side effects. Murphy emphasized that this study, like all genetic association studies, will need to be replicated, and the results may not apply to other antidepressants. The researchers will further analyze the data they have from this study, looking at more genes to see how they relate to antidepressant efficacy and side effect frequency. "We have many other markers in the pipeline in other candidate genes that we are exploring and so we feel like this is just sort of the first stab," said Murphy.
As a first stab, I wonder if people realize just how trivial this information is, in the greater scheme of comprehending how the body works. They looked at two genes, and found an experiential correlation in one of two genes, but that effect was a difference between 16% and 46% of subjects discontinuing due to intolerable side effects. People want to think that genes are 100% or 0% of an effect, not wishy-washy middle of the road numbers like 16% or 46%. Moreover, we aren't told if the discontinuers discontinued for the same reason, or numerous different ones.
There's been a lot of hype about the Human Genome Project (HGP). About how we'll soon understand everything about people, how miraculous treatments are at hand.....Not!
We're mapping the DNA of *one* person. The excerpt I've retained above describes 40 (known) variants of but one gene. Mapping one individual isn't going to tell us a friggin thing about gene variants. We don't even have an idea how genes vary among races. Up until now, most genes are sequenced in white people, and Japanese. There are lots of other people out there. We won't know we've got every gene variant until we sequence every person, not just one white guy.
One outcome of the HGP is the novel idea that there may only be about 30,000 genes in all our DNA (I think there are more, but there certainly can't be less). And I know there are many genes with huge numbers of variants (40 known versions of the CYP2D6 enzyme, for example), but let's settle on an average of six for each gene.
Okay, so how many different people can you construct, if there was a unique individual with every possible combination of those genes? That would 30,000 to the sixth power, or 24,300,000,000,000,000,000,000 individuals. Now, few diseases arise from problems in single genes (diseases like Tay-Sachs and Huntingdon's were identified simply because they involve one gene variant.....those are easy). What about diseases that arise from the interaction of two genes? To consider the possible ways that each of two genes might interact, in all those people, would require consideration of more interactions than there are grains of sand on Earth. If you consider three-gene and four-gene interactions, you're probably dealing with more grains of sand than there are in the Milky Way.
We're also commonly assuming that genes are on or off, or that they're dominant or recessive. Some cases, like eye colour, quite neatly fall into those cute little mental boxes. But that's not the way the vast majority of genes work. They're sort of on, or sort of off. Geneticists call that incomplete penetrance. Factor that into the mathematical possibilities. Just try.
Now factor in disturbances in DNA caused by viruses. And gamma rays from the sun. Environmental effects like exposure to stress, or pathogens, and dietary effects.
My mind boggles. And I hope yours does too.
To back off from what must seem to be a fatalistic viewpoint.....
What we learn changes our ability to guess correctly. A lot of what we "know" is not reproducible in text or mathematical formulae. That's what makes one doctor better than another, methinks. Or one musician. Etc.
I don't know that scientific reductionism is the way to go.....this gene here predicts that effect there 46% of the time.....what do we get from that, really? We already know that certain drugs are safer for geriatric use than others, ya know? Do we need the details?
I don't wanna burst anybody's bubble, but sometimes pin pricks are a part of life too. Like the guy who thought the $500 dollar test was gonna give him insight into what drug to take. Sorry, but, the only thing I can predict from that lab test is that lab will get your $500. Next!
Just my perhaps all too cynical but typically opinionated response.
Lar
poster:Larry Hoover
thread:295124
URL: http://www.dr-bob.org/babble/20031231/msgs/296015.html