[HN Gopher] A kids' novel inspired me to simulate a gene drive o...
___________________________________________________________________
A kids' novel inspired me to simulate a gene drive on 86M genealogy
profiles
Author : popcalc
Score : 49 points
Date : 2024-03-03 15:38 UTC (7 hours ago)
(HTM) web link (worthdoingbadly.com)
(TXT) w3m dump (worthdoingbadly.com)
| baq wrote:
| > For example, take the simple family tree above. If I'm person
| 28 on this family tree, and I'm making a genealogy, I would care
| about my own ancestors, but I might not care much about my
| ancestors' siblings (my aunts and uncles) and their descendants.
| So most of the family tree won't be present in a genealogy
| database
|
| Or, the data is correct and a lot of lines of people just die out
| without offspring.
| yorwba wrote:
| Or most people can't trace their lineage all the way back to
| 1809 and some ancestors just spawn into existence (or descend
| parthenogenetically from a single parent) at a later date.
|
| Maybe those missing ancestors could be imputed by sampling a
| random other individual with known parents born in the same
| year.
| dhosek wrote:
| I think that's likely it. One should instead build the tree
| backwards starting with the most recent members of the
| database and then find their earliest known ancestors. These
| can then be assigned the magic trait based on the likelihood
| of it being present based on their birth year cohort (or more
| reasonably birth decade or some other larger bucket).
| gwern wrote:
| No, that can't explain his result showing the witch gene going
| extinct despite, by stipulation, zero fitness impact on
| individuals but 100% transmission. Because the human population
| did grow over that time, so every deadend must have been
| balanced out by even more children by the non-deadends. And
| since he starts at hefty fractions like '1% of the entire
| population' (and goes up to 20%), we should be well beyond any
| issues of very small number effects or gambler's ruin.
|
| There has to be something wrong with either the data or code.
| (It almost looks like a sign error!)
| 082349872349872 wrote:
| > _because nobody wants to read fancode._
|
| for large enough values of nobody (nopony?), anyway.
___________________________________________________________________
(page generated 2024-03-03 23:00 UTC)