4 ms·

There was a paper not long ago saying that the DNA half life is about 500k years or so and that most of the DNA would be degraded. The issue I had with this is

There was a paper not long ago saying that the DNA half life is about 500k years or so and that most of the DNA would be degraded. The issue I had with this is why can't we reassemble chunks of DNA like we can with a puzzle. I'm certain specialized algorithms exist within DNA analysis for this reason. This isn't my field and there are already algorithms that exist on the top of my head that could solve this issue. The main issue is that you'd have to find a LOT of DNA to build one complete strand and you wouldn't have one strand from one specific animal. ... but you'd have a rough picture.

daemonk 7y ago
That's pretty much how modern genome assemblies work. We currently do not have a reliable or accurate way of sequencing DNA longer than ~200 bases. So we fragment the DNA into small pieces, sequence them and then put it all back together. It's not a completely solved problem as there are features of genomes that can make this process difficult or complicated (repetitive regions, highly heterozygous organisms, etc). Especially with short sequences. We have technology right now that can allow us to sequence long fragments, but at lower quality and accuracy. There are a lot of tools out there that uses the longer, but lower quality sequences to scaffold the shorter, higher quality sequencing data.
- haihaibye 7y ago
  Illumina sequencers (highest throughput per $) have a (paired end total) read length of 300 but we can sequence much longer with other tech. Ie Pacbio - 10-15kb Sanger sequencing, used for the human genome project 20 years ago is over 500 bases
jessriedel 7y ago
That's the halflife of each bond between nucleotides. After one tenth of a half life you're left with strands of average length ~20 nucleotides. After a full half life they are of length ~2, i.e., mostly just scattered GCs, ATs, etc. After several half-lives, it's almostly completely single nucleotides, and the chance that there are any surviving strands longer than a given length N falls exponentially in time and N.
Robotbeat 7y ago
What does half life of 500,000 years mean? Does that mean any given gene will, on average, split every 500,000 years? Or become unusable in 500,000 years? 65 million years would mean 130 halvings. Basically nothing left, then. ...but this may depend on environmental conditions. this must depend on average environmental temperature, right? Might dinosaur specimens in Antarctic or Arctic regions have experienced significantly lower average environmental temperatures? That could increase the half life significantly. Maybe we'll find Cryolophosaurus DNA some day... EDIT: the DNA degradation study specimen was in the south island of New Zealand, which had an average burial temperature of 13C. Antarctica along the coast has an average annual temperature of about -10C. Further inland and at higher elevation, it's much colder and dryer. Both those increase DNA stability. It's not crazy to me that you might have a factor of 100x greater stability for specimens found where the average temperature over the last 65 million years was 20-40 degrees colder and also dryer. EDIT AGAIN: Just to support what I said, it seems like a difference of average burial temperature of just 2.5 degrees C doubles the half-life of DNA (and this is multiplicative), so with a 20 degree difference in average burial temperature, that would be 8 doublings of half-life length (factor of 256), so instead of 500,000 years until the DNA was all broken up, you could have 128 million years. https://figshare.com/articles/_Predicted_DNA_half_life_for_varying_average_soil_temperature_for_each_fragment_type_/1420162 https://figshare.com/articles/_Predicted_DNA_half_life_for_v...
- gnramires 7y ago
  Not sure where his 500k years figure comes from. A cursory search gives 521 years: https://www.nature.com/news/dna-has-a-521-year-half-life-1.11555 https://www.nature.com/news/dna-has-a-521-year-half-life-1.1... Chemical decay tends to follow Arrhenius equation: https://en.wikipedia.org/wiki/Arrhenius_equation#Equation https://en.wikipedia.org/wiki/Arrhenius_equation#Equation so indeed a difference of maybe 40C should lead to a difference in half life of about 20-30%. I think about 30 half lives (~10^-9 degradation factor) should be the limit of recoverability, so perhaps 521x30 = 15k years in normal conditions up to 521x30x1.25 = 20k years in low temperature. Of course, other factors such as humidity could contribute as well to the half-life, and reactivity is a lot more complicated than the Arrhenius model in reality. But even then I would be surprised such a vast difference in reactivity outside of true cryogenic conditions (maybe even shielding from radiation?). edit: your second edit is interesting :) is that curve polynomial or exponential?
gorgoiler 7y ago
> reassemble like a puzzle That’s a good enough analogy for modern DNA sequencing. The code of DNA is an image trapped in an opaque box. Instead of being able to image the complete picture of an organism’s DNA, you take thousands of identical images in opaque boxes, smash the boxes and their contents into tiny pieces to get bits of image out, then line up the now exposed fragments to see the original image. It is the realignment that is the computationally expensive part of the process.
refurb 7y ago
The reason you can’t reassemble is because the puzzle is cut up so much it’s just individual bases. All of the sequence information is lost.

daemonk 7y ago

That's pretty much how modern genome assemblies work. We currently do not have a reliable or accurate way of sequencing DNA longer than ~200 bases. So we fragment the DNA into small pieces, sequence them and then put it all back together. It's not a completely solved problem as there are features of genomes that can make this process difficult or complicated (repetitive regions, highly heterozygous organisms, etc). Especially with short sequences. We have technology right now that can allow us to sequence long fragments, but at lower quality and accuracy. There are a lot of tools out there that uses the longer, but lower quality sequences to scaffold the shorter, higher quality sequencing data.

haihaibye 7y ago
Illumina sequencers (highest throughput per $) have a (paired end total) read length of 300 but we can sequence much longer with other tech. Ie Pacbio - 10-15kb Sanger sequencing, used for the human genome project 20 years ago is over 500 bases

jessriedel 7y ago

That's the halflife of each bond between nucleotides. After one tenth of a half life you're left with strands of average length ~20 nucleotides. After a full half life they are of length ~2, i.e., mostly just scattered GCs, ATs, etc. After several half-lives, it's almostly completely single nucleotides, and the chance that there are any surviving strands longer than a given length N falls exponentially in time and N.

Robotbeat 7y ago

What does half life of 500,000 years mean? Does that mean any given gene will, on average, split every 500,000 years? Or become unusable in 500,000 years? 65 million years would mean 130 halvings. Basically nothing left, then. ...but this may depend on environmental conditions. this must depend on average environmental temperature, right? Might dinosaur specimens in Antarctic or Arctic regions have experienced significantly lower average environmental temperatures? That could increase the half life significantly. Maybe we'll find Cryolophosaurus DNA some day... EDIT: the DNA degradation study specimen was in the south island of New Zealand, which had an average burial temperature of 13C. Antarctica along the coast has an average annual temperature of about -10C. Further inland and at higher elevation, it's much colder and dryer. Both those increase DNA stability. It's not crazy to me that you might have a factor of 100x greater stability for specimens found where the average temperature over the last 65 million years was 20-40 degrees colder and also dryer. EDIT AGAIN: Just to support what I said, it seems like a difference of average burial temperature of just 2.5 degrees C doubles the half-life of DNA (and this is multiplicative), so with a 20 degree difference in average burial temperature, that would be 8 doublings of half-life length (factor of 256), so instead of 500,000 years until the DNA was all broken up, you could have 128 million years. https://figshare.com/articles/_Predicted_DNA_half_life_for_varying_average_soil_temperature_for_each_fragment_type_/1420162 https://figshare.com/articles/_Predicted_DNA_half_life_for_v...

gnramires 7y ago
Not sure where his 500k years figure comes from. A cursory search gives 521 years: https://www.nature.com/news/dna-has-a-521-year-half-life-1.11555 https://www.nature.com/news/dna-has-a-521-year-half-life-1.1... Chemical decay tends to follow Arrhenius equation: https://en.wikipedia.org/wiki/Arrhenius_equation#Equation https://en.wikipedia.org/wiki/Arrhenius_equation#Equation so indeed a difference of maybe 40C should lead to a difference in half life of about 20-30%. I think about 30 half lives (~10^-9 degradation factor) should be the limit of recoverability, so perhaps 521x30 = 15k years in normal conditions up to 521x30x1.25 = 20k years in low temperature. Of course, other factors such as humidity could contribute as well to the half-life, and reactivity is a lot more complicated than the Arrhenius model in reality. But even then I would be surprised such a vast difference in reactivity outside of true cryogenic conditions (maybe even shielding from radiation?). edit: your second edit is interesting :) is that curve polynomial or exponential?

gorgoiler 7y ago

> reassemble like a puzzle That’s a good enough analogy for modern DNA sequencing. The code of DNA is an image trapped in an opaque box. Instead of being able to image the complete picture of an organism’s DNA, you take thousands of identical images in opaque boxes, smash the boxes and their contents into tiny pieces to get bits of image out, then line up the now exposed fragments to see the original image. It is the realignment that is the computationally expensive part of the process.

refurb 7y ago

The reason you can’t reassemble is because the puzzle is cut up so much it’s just individual bases. All of the sequence information is lost.