The Violinist's Thumb: And Other Lost Tales of Love, War, and Genius, as Written by Our Genetic Code
Page 33
Of course, scientists still need to analyze the bajillions of A’s, C’s, G’s, and T’s they’re gathering. Having been humbled by the HGP, they know they can’t just stare at the stream of raw data and expect insights to pop out, Matrix style. They need to consider how cells splice DNA and add epigenetic marginalia, much more complicated processes. They need to study how genes work in groups and how DNA packages itself in three dimensions inside the nucleus. Equally important, they need to determine how culture—itself a partial product of DNA—bends back and influences genetic evolution. Indeed, some scientists argue that the feedback loop between DNA and culture has not only influenced but outright dominated human evolution over the past sixty thousand years or so. Getting a handle on all of this will require serious computing horsepower. Craig Venter demanded a supercomputer, but geneticists in the future might need to turn to DNA itself, and develop tools based on its amazing computational powers.
On the software side of things, so-called genetic algorithms can help solve complicated problems by harnessing the power of evolution. In short, genetic algorithms treat the computer commands that programmers string together as individual “genes” strung together to make digital “chromosomes.” The programmer might start with a dozen different programs to test. He encodes the gene-commands in each one as binary 0s and 1s and strings them together into one long, chromosome-like sequence (0001010111011101010…). Then comes the fun part. The programmer runs each program, evaluates it, and orders the best programs to “cross over”—to exchange strings of 0s and 1s, just like chromosomes exchange DNA. Next the programmer runs these hybrid programs and evaluates them. At this point the best cross over and exchange more 0s and 1s. The process then repeats, and continues again, and again, allowing the programs to evolve. Occasional “mutations”—flipping 0s to 1s, or vice versa—add more variety. Overall, genetic algorithms combine the best “genes” of many different programs into one near-optimal one. Even if you start with moronic programs, genetic evolution improves them automatically and zooms in on better ones.
On the hardware (or “wetware”) side of things, DNA could someday replace or augment silicon transistors and physically perform calculations. In one famous demonstration, a scientist used DNA to solve the classic traveling salesman problem. (In this brainteaser, a salesman has to travel to, say, eight cities scattered all over a map. He must visit each city once, but once he leaves a city he cannot visit it again, even just to pass through on his way somewhere else. Unfortunately, the cities have convoluted roads between them, so it’s not obvious in what order to visit.)
To see how DNA could possibly solve this problem, consider a hypothetical example. First thing, you’d make two sets of DNA snippets. All are single-stranded. The first set consists of the eight cities to visit, and these snippets can be random A-C-G-T strings: Sioux Falls might be AGCTACAT, Kalamazoo TCGACAAT. For the second set, use the map. Every road between two cities gets a DNA snippet. However—here’s the key—instead of making these snippets random, you do something clever. Say Highway 1 starts in Sioux Falls and ends in Kalamazoo. If you make the first half of the highway’s snippet the A/T and C/G complement of half of Sioux Falls’s letters, and make the second half of the highway’s snippet the A/T and C/G complement of half of Kalamazoo’s letters, then Highway 1’s snippet can physically link the two cities:
After encoding every other road and city in a similar way, the calculation begins. You mix a pinch of all these DNA snippets in a test tube, and presto change-o, one good shake computes the answer: somewhere in the vial will be a longer string of (now) double-stranded DNA, with the eight cities along one strand, in the order the salesman should visit, and all the connecting roads on the complementary strand.
Of course, that answer will be written in the biological equivalent of machine code (GCGAGACGTACGAATCC…) and will need deciphering. And while the test tube contains many copies of the correct answer, free-floating DNA is unruly, and the tube also contains trillions of wrong solutions—solutions that skipped cities or looped back and forth endlessly between two cities. Moreover, isolating the answer requires a tedious week of purifying the right DNA string in the lab. So, yeah, DNA computing isn’t ready for Jeopardy. Still, you can understand the buzz. One gram of DNA can store the equivalent of a trillion CDs, which makes our laptops look like the gymnasium-sized behemoths of yesteryear. Plus, these “DNA transistors” can work on multiple calculations simultaneously much more easily than silicon circuits. Perhaps best of all, DNA transistors can assemble and copy themselves at little cost.
If deoxyribonucleic acid can indeed replace silicon in computers, geneticists would effectively be using DNA to analyze its own habits and history. DNA can already recognize itself; that’s how its strands bond together. So DNA computers would give the molecule another modest level of reflexivity and self-awareness. DNA computers could even help DNA refine itself and improve its own function. (Makes you wonder who’s in charge…)
And what kinds of DNA improvements might DNA computing bring about? Most obviously, we could eradicate the subtle malfunctions and stutters that lead to many genetic diseases. This controlled evolution would finally allow us to circumvent the grim waste of natural selection, which requires that the many be born with genetic flaws simply so that the few can advance incrementally. We might also improve our daily health, cinching our stomachs in by engineering a gene to burn high-fructose corn syrup (a modern answer to the ancient apoE meat-eating gene). More wildly, we could possibly reprogram our fingerprints or hairstyles. If global temperatures climb and climb, we might want to increase our surface area somehow to radiate heat, since squatter bodies retain more heat. (There’s a reason Neanderthals in Ice Age Europe had beer keg chests.) Furthermore, some thinkers suggest making DNA adjustments not by tweaking existing genes but by putting updates on an extra pair of chromosomes and inserting them into embryos*—a software patch. This might prevent intergenerational breeding but would bring us back in line with the primate norm of forty-eight.
These changes could make human DNA worldwide even more alike than it is now. If we tinker with our hair and eye color and figures, we might end up looking alike, too. But based on the historical pattern with other technologies, things might well go the other way instead: our DNA could become as diverse as our taste in clothing, music, and food. In that case, DNA could go all postmodern on us, and the very notion of a standard human genome could disappear. The genomic text would become a palimpsest, endlessly overwriteable, and the metaphor of DNA as “the” blueprint or “the” book of life would no longer hold.
Not that it ever really did hold, outside our imaginations. Unlike books and blueprints, both human creations, DNA has no fixed or deliberate meaning. Or rather, it has only the meaning we infuse it with. For this reason we should interpret DNA cautiously, less like prose, more like the complicated and solemn utterances of an oracle.
As with scientists studying DNA, pilgrims to the Delphic oracle in ancient Greece always learned something profound about themselves when they inquired of it—but rarely what they assumed they’d learned at first. The general-king Croesus once asked Delphi if he should engage another emperor in battle. The oracle answered, “You will destroy a great empire.” Croesus did—his own. The oracle informed Socrates that “no one is wiser.” Socrates doubted this, until he’d canvassed and interrogated all the reputedly wise men around. He then realized that, unlike them, he at least admitted his ignorance and didn’t fool himself into “knowing” things he didn’t. In both cases, the truth emerged only with time, with reflection, when people had gathered all the facts and could parse the ambiguities. The same with DNA: it all too often tells us what we want to hear, and any dramatist could learn a thing about irony from it.
Unlike Delphi, our oracle still speaks. From so humble a beginning, despite swerves and near extinctions, our DNA (and RNA and other ’NAs) did manage to create us—creatures bright enough to discover and decipher the DNA inside them. But
bright enough as well to realize how much that DNA limits them. DNA has revealed a trove of stories about our past that we thought we’d lost forever, and it has endowed us with sufficient brains and curiosity to keep mining that trove for centuries more. And despite that push-pull, gotta-have-it-won’t-stand-for-it ambivalence, the more we learn, the more tempting, even desirable, it seems to change that DNA. DNA endowed us with imagination, and we can now imagine freeing ourselves from the hard and heartbreaking shackles it puts on life. We can imagine remaking our very chemical essences; we can imagine remaking life as we know it. This oracular molecule seems to promise that if we just keep pushing, keep exploring and sounding out and tinkering with our genetic material, then life as we know it will cease. And beyond all the intrinsic beauty of genetics and all the sobering insights and all the unexpected laughs that it provides, it’s that promise that keeps drawing us back to it, to learn more and more and yet still more about our DNA and our genes, our genes and our DNA.
Epilogue
Genomics Gets Personal
Although they know better, many people versed in science, even many scientists, still fear their genes on some subliminal level. Because no matter how well you understand things intellectually, and no matter how many counterexamples turn up, it’s still hard to accept that having a DNA signature for a disease doesn’t condemn you to develop the disease itself. Even when this registers in the brain, the gut resists. This discord explains why memories of his Alzheimer’s-ridden grandmother convinced James Watson to suppress his apoE status. It also explains, when I plumbed my own genes, why boyhood memories of fleeing from my grandfather convinced me to conceal any hints about Parkinson’s disease.
During the writing of this book, however, I discovered that Craig Venter had published everything about his genome, uncensored. Even if releasing it publicly seemed foolhardy, I admired his aplomb in facing down his DNA. His example fortified me, and every day that passed, the discrepancy between what I’d concluded (that people should indeed face down their genes) and how I was behaving (hiding from my Parkinson’s status) nagged me more and more. So eventually I sucked it up, logged on to the testing company, and clicked to break the electronic seal on that result.
Admittedly, it took another few seconds before I could look up from my lap to the screen. As soon as I did, I felt a narcotic of relief flood through me. I felt my shoulders and limbs unwind: according to the company, I had no increased risk for Parkinson’s after all.
I whooped. I rejoiced—but should I have? There was a definite irony in my happiness. Genes don’t deal in certainties; they deal in probabilities. That was my mantra before I peeked, my way of convincing myself that even if I had the risky DNA, it wouldn’t inevitably ravage my brain. But when things looked less grim suddenly, I happily dispensed with uncertainty, happily ignored the fact that lower-risk DNA doesn’t mean I’ve inevitably escaped anything. Genes deal in probabilities, and some probability still existed. I knew this—and for all that, my relief was no less real. It’s the paradox of personal genetics.
Over the next months, I shooed away this inconvenient little cognitive dissonance and concentrated on finishing the book, forgetting that DNA always gets the last word. On the day I dotted the last i, the testing company announced some updates to old results, based on new scientific studies. I pulled up my browser and started scrolling. I’d seen previous rounds of updates before, and in each case the new results had merely corroborated what I’d already learned; my risks for things had certainly never changed much. So I barely hesitated when I saw an update for Parkinson’s. Fortified and foolhardy, I clicked right through.
Before my mind registered anything, my eyes lit on some green letters in a large font, which reinforced my complacency. (Only red lettering would have meant watch out.) So I had to read the accompanying text a few times before I grasped it: “Slightly higher odds of developing Parkinson’s disease.”
Higher? I looked closer. A new study had scrutinized DNA at a different spot in the genome from the results I’d seen before. Most Caucasian people like me have either CT or TT at the spot in question, on chromosome four. I had (per the fat green letters) CC there. Which meant, said the study, higher odds.
I’d been double-crossed. To expect a genetic condemnation and receive it in due course is one thing. But to expect a condemnation, get pardoned, and find myself condemned again? Infinitely more torture.
Somehow, though, receiving this genetic sentence didn’t tighten my throat as it should have. I felt no panic, either, no fight-or-flight jolt of neurotransmitters. Psychologically, this should have been the worst possible thing to endure—and yet my mind hadn’t erupted. I wasn’t exactly pumped up about the news, but I felt more or less tranquil, untroubled.
So what happened between the first revelation and the second, the setup and the would-be knockdown? Without sounding too pompous, I guess I got an education. I knew now that for a complex disease like Parkinson’s—subject to the sway of many genes—any one gene probably contributes little to my risk. I then investigated what a “slightly higher” risk meant, anyway—just 20 percent, it turns out. And that’s for a disease that affects (as further digging revealed) only 1.6 percent of men anyway. The new study was also, the company admitted, “preliminary,” subject to amendments and perhaps outright reversals. I might still be saddled with Parkinson’s as an old man; but somewhere in the generational shuffling of genes, somewhere between Grandpa Kean and Gene and Jean, the dangerous bits might well have been dealt out—and even if they’re still lurking, there’s no guarantee they’ll flare up. There’s no reason for the little boy in me to keep fleeing.
It had finally penetrated my skull: probabilities, not certainties. I’m not saying personal genetics is useless. I’m glad to know, for instance (as other studies tell me), that I face higher odds of developing prostate cancer, so I can always make sure the doctor dons a rubber glove to check for that as I age. (Something to look forward to.) But in the clinic, for a patient, genes are just another tool, like blood work or urinalysis or family history. Indeed, the most profound changes that genetic science brings about likely won’t be instant diagnoses or medicinal panaceas but mental and spiritual enrichment—a more expansive sense of who we humans are, existentially, and how we fit with other life on earth. I enjoyed having my DNA sequenced and would do it again, but not because I might gain a health advantage. It’s more that I’m glad I was here, am here, in the beginning.
NOTES AND ERRATA
Chapter 1: Genes, Freaks, DNA
The 3:1 ratio: Welcome to the endnotes! Wherever you see an asterisk (*) in the text, you can flip back here to find digressions, discussions, scuttlebutt, and errata about the subject at hand. If you want to flip back immediately for each note, go right ahead; or if you prefer, you can wait and read all the notes after finishing each chapter, as a sort of afterword. This first endnote provides a refresher for Mendelian ratios, so if you’re comfortable with that, feel free to move along. But do flip back again. The notes get more salacious. Promise.
The refresher: Mendel worked with dominant traits (like tallness, capital A) and recessive traits (like shortness, lowercase a). Any plant or animal has two copies of each gene, one from Mom, one from Dad. So when Mendel crossed AA plants with aa plants (below, left), the progeny were all Aa and therefore all tall (since A dominates a):
| A | A | | A | a |
a | Aa | Aa | A | AA | Aa |
a | Aa | Aa | a | Aa | aa |
When Mendel crossed one Aa with another (above, right), things got more interesting. Each Aa can pass down A or a, so there are four possibilities for the offspring: AA, Aa, aA, and aa. The first three are again tall, but the last one will be short, though it came from tall parents. Hence a 3:1 ratio. And just to be clear, the ratio holds in plants, animals, whatever; there’s nothing special about peas.
The other standard Mendelian ratio comes about when Aa mates with aa. In this case, half the children will be aa and won’t show
the dominant trait. Half will be Aa and will show it.
| A | a |
a | Aa | aa |
a | Aa | aa |
This 1:1 pattern is especially common in family trees when a dominant A trait is rare or arises spontaneously through a mutation, since every rare Aa would have to mate with the more common aa.
Overall the 3:1 and 1:1 ratios pop up again and again in classic genetics. If you’re curious, scientists identified the first recessive gene in humans in 1902, for a disorder that turned urine black. Three years later, they pinned down the first dominant human gene, for excessively stubby fingers.
Chapter 2: The Near Death of Darwin
until the situation blew over: The details about Bridges’s private life appear in Lords of the Fly, by Robert Kohler.
to proceed by jumps: When they were both young, in the 1830s, Darwin convinced his first cousin, Francis Galton, to drop out of medical school and take up mathematics instead. Darwin’s later defenders must have rued this advice many, many times, for it was Galton’s pioneering statistical work on bell curves—and Galton’s relentless arguments, based on that work—that most seriously undermined Darwin’s reputation.