There seems to be a fractal quality to the concept of genomic medicine at present: we get closer and closer to identifying potentially harmful alleles, but the more data we get, the more work remains to be done. Analytic techniques keep improving, but the prospects of translating data into operational therapeutic interventions remain murky.
The difficulty seems to lie in the basic concept of the human genome. Nature's news section summed up recent discoveries nicely in a headline:
Humans riddled with rare genetic variants
Each variant is itself unusual; and every genome has lots of them. One of two papers published simultaneously in Science recently (Nelson et al., a team led by John Novembre and Vincent Mooser) reports the results of sequencing 202 genes in 14,002 individuals:
We observed on average one variant per 17 bp [base pairs] in the overall sample and one variant per 21 bp in the Europeans (table S5). Among all variants, more than 95% were rare… and more than 74% were observed in only one or two subjects. ~90% of rare variants were not previously reported, as opposed to ~5% of common variants.
The other paper (Tennessen et al., led by Joshua Akey and Michael Bamshad) produced similar results from another large sample:
~95.7% of SNVs [single nucleotide variants] predicted to be functionally important were rare.
What accounts for this unexpected finding? In part it may be due to the human population having grown "by at least three orders of magnitude over the past 400 generations," as noted in a third recent Science paper, by Alon Keinan and Andrew Clark. (That's only about 11 years of Drosophila reproduction, though fruit flies do have many more individuals per generation.) This "increases the load of rare variants and is likely to play a role in the individual genetic burden of complex disease risk." Basically, we have been moving away from genetic equilibrium and haven't had time for natural selection to work thoroughly.
The New York Times had a sober take on these results:
The task of finding the genetic roots of common disease seems much harder, dimming the promise of personal genomics and the chances of quick medical payoffs from the human genome project ... Until now, researchers have looked only at common mutations as possible contributors to the risk of common diseases ... the "common disease, common variant" hypothesis ... But the theory was wrong. Common variants have turned out to explain only a fraction of the genetic risk of common disease.
Meanwhile, there has been increasingly contentious discussion about the possible costs of interpreting genomic data. A report in GEN clearly overstated them (one of the authors essentially retracted it) and provoked a useful rebuttal by Daniel MacArthur, and his commenters. But the complexity of analysis is real, and costs could easily reach the point where, to quote MacArthur, "you're going to need some very good insurance indeed." And then there's the point that its utility is yet to be proven.
There is a possible exception: predicting a patient's response to specific drug treatments. Drug sensitivity is becoming a major research area for genomics, for instance in connection with cancer treatments. It is probably no coincidence that 15 of the 31 co-authors in the Novembre-Mooser team work for GlaxoSmithKline, and conclude their paper:
Because the genes studied are related to drug discovery, development or repositioning efforts this work has potential to help investigate target biology and drug response.
No one is giving up on genomic medicine, that's clear. The concept of the variable human genome is now taking enough hold that there is talk of healthy people establishing their own "personal reference genome" (there's a new, expanded job market for sequence salespeople). Everyone involved seems to be sure that something useful will come of all this. But it's hard to see the path they seem to be sure is there.
And one final thought: If we cannot even define a healthy genome precisely, can we please put away fantasies of generating a superhuman one?
Previously on Biopolitical Times:
Posted in Genetic Selection, Personal genomics, Pete Shanks's Blog Posts, Sequencing & Genomics
Comments are now closed for this item.