Hacker News Reader

Comment by tristanj | Hacker News Reader

Genetic engineering can't bring back most extinct animals. DNA molecules have a half life of around ~500 years and samples older than a few hundred thousand years will have too many errors to recreate. Better statistical techniques will push this bound but it's hard to argue against physics. It's estimated that DNA older than 20 million years old will be fully unrecoverable.

Notice how all the author's examples are recently extinct species, not ancient ones. I guess they don't mention this because it doesn't fit with their worldview.

Some sources:

http://www.the-scientist.com/?articles.view/articleNo/32799/...

https://en.wikipedia.org/wiki/Ancient_DNA

View on HN

Could you not produce viable samples by collecting multiple copies of the DNA and overlapping them, thus removing the errors that would likely happen at random locations?

Not likely. Dude to the chemistry of nucleotides, they can degrade in a particular pattern that obscures the ancestral allele. For example the deamination of cytosine to thymine can leave us with an ambiguity in what in our read for the correct nucleotide is.

Was it thymine? Or was it cytosine that degraded to thymine? That means for literally 50% of the genome we have a huge ambiguity and few ways to tell what it really is.

We can use the ancestral allele from a nearby extant species, but then we miss out on all the extinct variation (which is the important stuff).

We can try to increase coverage of the ancient genome to capture loci where cytosines haven't deaminated, but thats tough for a dozen reasons, and in the end you'll still have no confidence in 25% of the genome.

We can get very close to a full ancient genome, but never 100%. Maybe through transgenics, selective breeding, and ancient sequencing we can re-create ancient species, but we will never make them exactly as they once were.

    we will never make them exactly as they once were

And maybe that could be good enough and better than not doing it at all...

> Better statistical techniques will push this bound

That's what he meant by statistical techniques, we don't currently read a full "stream" even with brand new DNA, we smash it to bits then read those and have the computers put them back together.

At a certain point though what's left is too garbled or in too small sequences so that there isn't enough overlap that you can't put humpty back together again :).

Yes, but with a half life of 500 years, the amount of information you have halves every 500 years you go back. At some point there just isn't enough information to go viably go back further.

Apart from very close species (like incubating passenger pigeon DNA in a fantail pigeon host), is it possible to recreate say a woolly mammoth even assuming you can perfectly reconstruct the DNA itself?

That is, how much of an organism is encoded in the DNA and how much in the mother's body? And yes, this question is derived from GEB[0].

[0] https://en.wikipedia.org/wiki/G%C3%B6del,_Escher,_Bach

Has there been much work on cloning endangered species? I guess resurrecting extinct species would be "cooler", but helping endangered species would be less controversial.

Reply on news.ycombinator.com