Extracting Family Trees From Ancient Genomes

New techniques and very old bones overcome the limits of genome sequencing for prehistoric horses, ancient cave bears, and even our own early ancestors.

For millennia, the stories of long-extinct species — including our own progenitors — have been buried with their skeletal remains. But in 2013, ultramodern DNA extraction and sequencing techniques enabled researchers to access ancient genetic codes and translate their evolutionary tales: Researchers in Denmark reconstructed a record-breaking 700,000-year-old horse genome, and geneticists in Germany began parsing the DNA of 400,000-year-old hominids.

Geologists saw the first glint of the horse’s history in 2003 when they plucked its toe bone from permafrost in a remote Yukon gold mine. The uninterrupted freeze of the permafrost preserved DNA in the horse bone, but since DNA decays into smaller and less intelligible fragments over time, the specimen seemed too ancient to analyze. “When that fossil was found, no one would have believed that we could get DNA out of it,” says Yukon government paleontologist Grant Zazula.

Armed with a decade of improvements in next-generation sequencing techniques, Denmark-based evolutionary geneticist Ludovic Orlando could finally piece together what was left of the bone’s DNA. Using what’s called true single molecule sequencing, Orlando lit up the A’s, C’s, T’s and G’s, one by one, to assemble the horse’s genome — six times older than any nuclear DNA specimen ever sequenced.

The results, published in July, radically revise the timeline for equine evolution, revealing that the common ancestor of contemporary horses, zebras and donkeys originated at least 4 million years ago, twice as far back
as previously thought.

Hunting for Human History

Deciphering human DNA of the same vintage seems like it should be next on the docket. Here’s the catch: No comparably ancient human skeleton has been found preserved in permafrost. Weathered hominid bones — and their decaying DNA — are generally discovered in temperate caves, like the one in Spain’s Atapuerca Mountains, whose collection of these remains is among the world’s largest and oldest. “This is a dream site for studying the ancestors of Neanderthals and perhaps modern humans,” says evolutionary geneticist Svante Pääbo.

But getting a DNA sample from a bone means drilling a hole in it, and archaeologists were not about to let geneticists go to work on the deteriorating human skeletons without some guarantee of a genome. So Pääbo’s team procured a similarly degraded non-human specimen from the same rocky dwelling for their proof of concept, published in August: the genome of a 400,000-year-old cave bear.

The Germany-based team developed two advances to get and use more of the bear’s genetic information from its bones. First, they salvaged DNA fragments degraded down to as few as 30 base pairs (by comparison, fragments from the frozen horse bone averaged 78 base pairs). Second, they separated the complementary strands of DNA in these fragments before sequencing so they could still use one half of the double helix even if the other half were damaged.

Now Pääbo’s team is applying these techniques to Atapuerca’s ancient hominids to pinpoint changes in the human genome and determine when they occurred. “If we can see things directly — things that were alive 400,000 years ago,” says Jesse Dabney, a doctoral student involved with the project, “we can get a clearer picture of our own evolution.”

Two Elusive Prime Number Problems Solved

After centuries of flummoxing number crunchers, two mathematical puzzles about prime numbers were cracked this year.

Prime numbers — those divisible only by 1 and the number itself, like 5, 11 or 37 — are like the atoms of mathematics: All numbers are formed by multiplying these building blocks together. 

But what happens when you add a number to a prime number? When will the sum be prime? Or, conversely, when is a number a sum of primes? Mathematicians have been working to answer these fundamental questions for centuries, and on the same day in May, two mathematicians finally found tantalizing partial answers to both of them.

To imagine the answer to the first question, start by adding the number 2 to a prime. When the sum is also prime, the pair is called a “twin prime,” like 5 and 7. As numbers get bigger, primes become more rare; you might then expect the spacing between them to grow consistently larger, too, so that very large twin primes would never occur.

Yet the famous but unproven “twin prime conjecture” states there are an infinite number of primes that differ by 2 — no matter how high you count, you will never run out of twin primes. A related, more general conjecture suggests there are also infinitely many pairs of primes that differ by 4, or 6, or any even number at all. 

But conjecture is all it was until May 13, when a nearly unknown mathematician, Yitang Zhang of the University of New Hampshire, made a serious dent in the twin primes conjecture. During a talk at Harvard, he presented a proof of the related, general conjecture that as prime numbers increase toward infinity, the spaces between them — counterintuitively — do not always do the same: No matter how big prime numbers get, you’ll always find pairs of them that differ by, at the very most, 70 million. 

Admittedly, 70 million is a lot bigger than 2, so the twin primes conjecture remains unsolved. But Zhang established for the first time a necessary (and supremely difficult) first step — that the spread between successive primes does not increase toward infinity.

On the same day Zhang emerged from obscurity to reveal his stunning proof, Harald Helfgott of the Ecole Normale Superieure in Paris cracked another famously elusive problem involving prime numbers — a variation on the Goldbach conjecture, which claims that every even number greater than 2 is the sum of two primes. (For example: 16 = 5 + 11.) 

Instead, Helfgott posted a proof of the “odd Goldbach conjecture,” which states that every odd number above 5 is the sum of three primes. (19 = 3 + 5 + 11.) It’s a big step in the right direction because the full Goldbach conjecture implies the odd version: Just take your odd number (say, 19), subtract the prime number 3 (now you have 16), and apply the Goldbach conjecture to the resulting even number. (16 = 5 + 11.)

While Helfgott’s proof does not solve the full conjecture, which is considered much harder, it shines a light on the intricate dance prime numbers engage in. Now the full conjecture, along with Zhang’s almost-but-not-quite-proven twin primes conjecture, remain a tantalizing plum for future mathematicians