The balance of experiment and theory is shifting in genomics; this matters for ELSI
This commentary in the Genomics Law Report’s ongoing series What ELSI is New? is contributed by Robert Cook-Deegan, Duke University Institute for Genome Sciences & Policy.
I am teaching Darwin this semester with a colleague from the English Department. It’s a real gas to go back to the roots of modern biology and see the shift from pure field observation and categorization to theoretically driven argument. Since Warren Weaver coined the term and then primed the pump for molecular biology in 1938 at the Rockefeller Foundation, biology has become a much more experimental science. Many lineages of embryology, genetics, and other parts of biology always were experimental, but after World War II, molecular approaches became dominant, driven in no small part by a succession of new technologies: ultracentrifuges, radio-isotopes incorporated into macromolecules, protein sequencing and synthesis, recombinant DNA, and nucleic acid sequencing. Sequencing really caught fire in the 1980s, and joined forces with the broader revolution in computing and communication using the Internet and computational firepower following Gordon Moore’s trajectory of faster, cheaper, and smaller computing power.
Many other parts of biology are changing too, and this is not a claim of genomic exceptionalism. It is merely a claim that things are changing, and fast, not that they are only changing in genomics (just think of stem cells). As I think about “what’s new,” I find myself drawn to one hugely important changethe log-a-year improvements in sequencing technology that suggest we’ll be generating one whale of a lot of DNA sequence data, and not just about Homo sapiens (I hope). I don’t pretend to know what that means. But I am drawn to two lines of research that seem to give us a glimpse of some major changes ahead that we would do well to be thinking about. I don’t quite know what to say about them, except to say that my antennae are quivering with a sense that they are really important.
One development is about what used to be called regulation of gene expression. It’s getting too complicated for that to really capture the meaning, but the general idea is that the genome is not even close to static, or the passive repository of Mendel’s particles of inheritance. The ENCODE project is throwing up all sorts of new insights, casting doubt on assumptions that conserved elements are selected for function and unconserved sequences are adrift—or that unexpressed regions can be called “junk.” Sydney Brenner used to joke that “junk” does not mean useless, but the stuff you keep in the attic. Seems to me like this “junk DNA” is all over the kitchen, family room, and not least the bedroom, but there’s not nearly so much up in the attic as we thought.
Why does this matter for ELSI? Let me count the ways, but start with just one to illustrate. On the Law axis of ELSI, we have 50,000 DNA patents largely premised on what we thought was true about the genome. Remember when we thought there were over 100,000 genes, or that one gene encoded one RNA encoding one protein? That’s so yesterday. This legacy of patents may begin to expire before the stakes are worth fighting over in court—or they may not. But one thing’s for sure, a cell is not a bag of genes that can be specified and patented under the Central Dogma of molecular biology. Yet parse the fractured English of many patent claims and you’ll see Watson, Crick, Beadle and Tatum laid bare.
The other tectonic movement in biology is beautifully illustrated by the paper from the Broad Institute on the human population genetic history of India by David Reich and his colleagues. It’s mind-blowing how off-the-shelf chips derived from common variations discovered in the HapMap can generate data on 125 people to reach powerful inferences using fancy math. Hypotheses about population relatedness can be tested and falsified. That’s using technology that began to take off a decade ago. Imagine now what we’ll be contending with as sequencing enters the fray. We should have an order or two more informative data, made even richer if we preserve the information about who passed genes to whom in the data sets. It seems quite likely that the granularity of human population genetics is poised to rocket because it won’t take all that many people to make inferences about any human lineage that we study.
I was once naive enough to think that the study of human population genetics would conquer racism, or at least strongly challenge it. I thought science might put a stake through the dark heart of eugenics and racial hygiene. That’s because I didn’t know much about racism except some of the genetics of human inheritance. Genomics won’t conquer racism any more than evolutionary biology has beaten back Creationism, and Maynard Olson, among others, in “Davenport’s Dream” has noted that we can’t count on the facts reinforcing every Liberal Dream. We may find some phenotype-genotype associations that we find downright uncomfortable that cannot be attributed to schlock science. It is not a given, but it is a possibility. I just have to think there are many, many issues that are going to arise as it becomes cheap and fast to study the full genomes of lots of people, and as we can reconstruct our common and our uncommon ancestry with levels of specificity vastly beyond what we have been able to do until now.
