4 ms·

Well, what half-life do you think that article implies? I took a shot at the calculation, and I got 1375 years.* That suggests to me that, yes, the popular figu

Well, what half-life do you think that article implies? I took a shot at the calculation, and I got 1375 years.* That suggests to me that, yes, the popular figure may be off, but what difference does it make when you're talking about dinosaurs? I mean it seems like you are implicitly saying that it's a huge exaggeration to claim 120,000 half-lives since the dinosaurs went extinct whereas it might be as little as 47,000. I don't think you're wrong. But raising 2 to either power overflows my calculator, so I don't see how it matters. *my calculation is very imprecise and ignorant of biology, but I think orders of magnitude are good enough when talking about exponentials. - the sample was 180 mg - the number of cells in that size sample might be 8e10.[1] - they had 17x coverage - log2(17/8e10) is about -32 - 44,000 years / 32 = 1375 [1] I got a number for cells/kg by assuming a human weighs 180 lbs and has the number of cells mentioned in this reference: https://pubmed.ncbi.nlm.nih.gov/23829164/ https://pubmed.ncbi.nlm.nih.gov/23829164/

credit_guy 6y ago
I don't think it works quite like that. A DNA molecule (some people prefer to call it a "strand") has billions of nucleotides (or "DNA letters"). There is one DNA molecule per chromosome, and a few dozens chromosomes in every single cell. For all practical purposes, the same chromosome in various cells has the exact same copy of the DNA molecule. After the organism's death, each DNA molecule will probably break down in a matter of days, at most. From this point of view a half-life of 521 years is an astronomic number, and totally false. After a few years there will be exactly zero intact DNA strands. But that won't make a difference anyway, as we are not able to read a full strand in one read anyway. Instead we read fragments and stitch them together. Like you see a "GATACA" and then a "TACACCG" and hypothesize that they were both parts of the longer string "GATACACCG". The longer the intersection string, the most likely the stitching is correct. In the mammoth case, they had fragments longer than 60 letters. Stitching together all these fragments, they were able to produce the full string (of billions of letters) 17 independent times. That gives them a high level of confidence that what they did is correct. In the case of dinosaurs. Will we ever be able to read fragments of tens of letters and stitch them together? Maybe, maybe not. If the DNA breaks down so much that you won't be able to see even pairs of letters, then we are totally out of luck. If we see pairs of letters, we probably see triplets, quadruplets, etc. Can we stitch together a long string out of quadruplets? Most likely not, but if we have 10-letter strings, probably yes. Are we going to have a contamination problem? If most of the DNA fragments come from bacteria rather than the original dinosaur, will we be doomed? Most likely we'll be able to come up algorithms that will eliminate bacterial DNA genes out of the reconstituted genome. We are quite smart. Is Jurassic Park in the near future. Not before we are able to revive mammoths. And a mammoth is difficult to create for the simple reason that it takes a very long time to gestate a baby elephant. If each iteration in your experiment takes one year, you are in danger of shipwrecking your career. Can we overcome these challenges? You bet. Anytime soon? My guess is that not during my lifetime, at least. But then Bill Gates states that we overestimate what we can do in one week but underestimate what we can do in a decade. So, you never know.
- dragonwriter 6y ago
  > A DNA molecule (some people prefer to call it a "strand") has billions of nucleotides (or "DNA letters"). A DNA molecule consists of two strands. RNA is single stranded.
- perl4ever 6y ago
  "I don't think it works quite like that" I'm sure it doesn't, but we agree that the recovery of the sequence implies a half life greater than 500 years, right? I was trying to get at how much greater, whether double or 100x or 1,000,000x, etc. "From this point of view a half-life of 521 years is an astronomic number, and totally false" Please set aside this point of view, because it wasn't mine. If they got 17x coverage, that's not literally from 17 totally intact cells, of course I know that. But I was interpreting the recoverable DNA in aggregate as being what the half-life refers to, as it were, proportional to 17 cells worth. If that's within a couple orders of magnitude, it's good enough for my point. If you think the half life could be millions of years, because I don't understand gene sequencing, then it would be really cool if you could do your own estimate and show me why your number is different.
  - credit_guy 6y ago
    I don't think the genome sequence coverage number can be used to infer anything about how fast DNA decays, or even how much DNA was available to the researchers. For example, from wikipedia [1], I found out that the African elephant's genome was sequenced in 2009; the details can be found in [2], where the sequencing coverage is listed at 7x. Why 7x and not 1000x, or more? Most likely there are some limiting factors, such as compute costs, and people are not chasing higher coverage numbers for their own sake. >If you think the half life could be millions of years I don't think the half life could be millions of years. I simply think half life is not a meaningful concept for DNA. Books decay. What's the half-life of books? How many books from a hundred years ago are still around nowadays? Let's be generous and say 0.1%, that's a half-life of about 10 years. That means there should be no books older than 300 years, yet I have seen with my own eyes a Gutenberg bible, printed well before 1500. [1] https://en.wikipedia.org/wiki/List_of_sequenced_animal_genomes#Mammals https://en.wikipedia.org/wiki/List_of_sequenced_animal_genom... [2] http://genome.ucsc.edu/cgi-bin/hgGateway?org=Elephant&db=loxAfr3&hgsid=154866831 http://genome.ucsc.edu/cgi-bin/hgGateway?org=Elephant&db=lox...

credit_guy 6y ago

I don't think it works quite like that. A DNA molecule (some people prefer to call it a "strand") has billions of nucleotides (or "DNA letters"). There is one DNA molecule per chromosome, and a few dozens chromosomes in every single cell. For all practical purposes, the same chromosome in various cells has the exact same copy of the DNA molecule. After the organism's death, each DNA molecule will probably break down in a matter of days, at most. From this point of view a half-life of 521 years is an astronomic number, and totally false. After a few years there will be exactly zero intact DNA strands. But that won't make a difference anyway, as we are not able to read a full strand in one read anyway. Instead we read fragments and stitch them together. Like you see a "GATACA" and then a "TACACCG" and hypothesize that they were both parts of the longer string "GATACACCG". The longer the intersection string, the most likely the stitching is correct. In the mammoth case, they had fragments longer than 60 letters. Stitching together all these fragments, they were able to produce the full string (of billions of letters) 17 independent times. That gives them a high level of confidence that what they did is correct. In the case of dinosaurs. Will we ever be able to read fragments of tens of letters and stitch them together? Maybe, maybe not. If the DNA breaks down so much that you won't be able to see even pairs of letters, then we are totally out of luck. If we see pairs of letters, we probably see triplets, quadruplets, etc. Can we stitch together a long string out of quadruplets? Most likely not, but if we have 10-letter strings, probably yes. Are we going to have a contamination problem? If most of the DNA fragments come from bacteria rather than the original dinosaur, will we be doomed? Most likely we'll be able to come up algorithms that will eliminate bacterial DNA genes out of the reconstituted genome. We are quite smart. Is Jurassic Park in the near future. Not before we are able to revive mammoths. And a mammoth is difficult to create for the simple reason that it takes a very long time to gestate a baby elephant. If each iteration in your experiment takes one year, you are in danger of shipwrecking your career. Can we overcome these challenges? You bet. Anytime soon? My guess is that not during my lifetime, at least. But then Bill Gates states that we overestimate what we can do in one week but underestimate what we can do in a decade. So, you never know.

dragonwriter 6y ago
> A DNA molecule (some people prefer to call it a "strand") has billions of nucleotides (or "DNA letters"). A DNA molecule consists of two strands. RNA is single stranded.
perl4ever 6y ago
"I don't think it works quite like that" I'm sure it doesn't, but we agree that the recovery of the sequence implies a half life greater than 500 years, right? I was trying to get at how much greater, whether double or 100x or 1,000,000x, etc. "From this point of view a half-life of 521 years is an astronomic number, and totally false" Please set aside this point of view, because it wasn't mine. If they got 17x coverage, that's not literally from 17 totally intact cells, of course I know that. But I was interpreting the recoverable DNA in aggregate as being what the half-life refers to, as it were, proportional to 17 cells worth. If that's within a couple orders of magnitude, it's good enough for my point. If you think the half life could be millions of years, because I don't understand gene sequencing, then it would be really cool if you could do your own estimate and show me why your number is different.
- credit_guy 6y ago
  I don't think the genome sequence coverage number can be used to infer anything about how fast DNA decays, or even how much DNA was available to the researchers. For example, from wikipedia [1], I found out that the African elephant's genome was sequenced in 2009; the details can be found in [2], where the sequencing coverage is listed at 7x. Why 7x and not 1000x, or more? Most likely there are some limiting factors, such as compute costs, and people are not chasing higher coverage numbers for their own sake. >If you think the half life could be millions of years I don't think the half life could be millions of years. I simply think half life is not a meaningful concept for DNA. Books decay. What's the half-life of books? How many books from a hundred years ago are still around nowadays? Let's be generous and say 0.1%, that's a half-life of about 10 years. That means there should be no books older than 300 years, yet I have seen with my own eyes a Gutenberg bible, printed well before 1500. [1] https://en.wikipedia.org/wiki/List_of_sequenced_animal_genomes#Mammals https://en.wikipedia.org/wiki/List_of_sequenced_animal_genom... [2] http://genome.ucsc.edu/cgi-bin/hgGateway?org=Elephant&db=loxAfr3&hgsid=154866831 http://genome.ucsc.edu/cgi-bin/hgGateway?org=Elephant&db=lox...