[ExI] crispr question

[Anders Sandberg]

On 2016-04-07 17:29, spike wrote:
>
> Cool, so this leads into my next comment.  We saw that it was 60 years 
> between the first marginal chess program to the orders-of-magnitude 
> more impressive champion AlphaGo.  We know that cheap technology 
> exists in which a prole can get her DNA file.  If a number of siblings 
> have that DNA file, they can figure out the genome of their parents, 
> and if several cousins compare their genomes, they can piece together 
> the genomes of their mutual grandparents, second cousins their great 
> grandparents and so on.
> If we could synthesize a DNA strand by some means using a DNA file, we 
> could perhaps use it to clone someone who lived a long time ago, given 
> sufficient numbers of their descendants have their DNA files.

Remember that each offspring gets 50% of the genome (lets ignore X and Y 
chromosomes). So if you have a sibling, you can reconstruct about 75% of 
either of your parents. If you have two, 87.5%, and so on. To get all 3 
billion base pairs right you need 31 siblings

Going back one generation means you have 25% of (say) grandfather. So 
now you need not just a bunch of siblings to build dad's genome (50% of 
grandfather), but you need to reconstruct more uncles and aunts (about 
31 of them) using 31 children each. This gets tricky fast, if your 
family does not breed like rabbits.

Accepting a bit less fidelity (say 90%) just requires two-three siblings 
for a parent, and two-three reconstructed uncles (about 6 or 7 starting 
people) for grandfather. But for the generation before that you need 
around 16 people, then 39, then 97... And in fact, 90% fidelity going 
back four generatons is just 65% fidelity.

The challenge is that each time we reproduce, half of the genome is 
discarded. Yes, we get a lovely replacement half from our partner, but 
information is lost. If things go well it is the less important 
information that is lost.


-- 
Anders Sandberg
Future of Humanity Institute
Oxford Martin School
Oxford University

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20160407/5ede3913/attachment.html>