5 ms·

I'm not a bioinformaticist. Try the 23andme algorithm, or whatever they used in the research that compared self-reported race to genetically measured ancestry.

I'm not a bioinformaticist. Try the 23andme algorithm, or whatever they used in the research that compared self-reported race to genetically measured ancestry. (The objective definition of your race is defined by where your ancestors come from, which, due to how babies get made, can be measured from your DNA. You can read more about this "DNA pseudo-science" here: https://en.wikipedia.org/wiki/Meiosis https://en.wikipedia.org/wiki/Meiosis ) College admissions departments already check to see if applicants are lying about race, e.g. by looking up their social media. Instead of scanning DNA, they could ask for some documentary evidence if they think an applicant is lying.

Mordisquitos 5y ago
I am a biologist and a bioinformatician, though not currently active in those fields. There is no such thing as an "objective definition of race", and claiming that so-called race can be determined from one's DNA is indeed pseudo-science.
- SamReidHughes 5y ago
  Then why can you tell what race people are by looking at them? I claimed you can predict 99.9999% of people's self-identified race from their DNA. Is that wrong? How high does that number have to be before you'd call it non-pseudo-science?
- carrionpigeon 5y ago
  > claiming that so-called race can be determined from one's DNA is indeed pseudo-science. That's taking it way too far. Genotyping populations identifies variation, and in cases where populations that have been reproductively isolated for tens of thousands of years (e.g. western Europeans and west/central Africans) are now in proximity, there will be substantial variation that is idiosyncratic to each population. In the US, reproductively isolated populations were suddenly placed in the same territory, and taboos on inter-racial mating persisted until very recently (50% toleration in the mid-1990s). One can accurately tell whether an American is phenotypically white versus black (or mixed) by DNA. That isn't to say that our ability to predict accurately black-vs-white will remain. The more race-mixing, the less ability to accurately categorize. We'll instead move to a situation similar to the one in which 'white' people find themselves today, where they are mixes of x percent of a variety of populations and not categorically 'Dutch,' 'Irish,' 'Finnish,' etc. Race is a sociological concept, yes, but it reflects reality in that people are visibly different. These differences are genetic in origin. It just so happens that in some cases, (e.g. two reproductively isolated and visibly distinct populations suddenly in contact with one another), these social categorizations will map to biological ones. In other cases where there is convergent evolution in terms of physical appearance, they won't.
notch656a 5y ago
Can you refer me to the 23andme algorithm that determines race with 99.9999% parity with self-reported race? Where do they report this outcome? I'm unable to determine the algorithm or find this claim by 23andme just from the reference to '23andme algorithm.' I've been totally unable to find such claims made by 23andme, suggesting they predict 'race' with any certainty to self-reported 'race.' I know what meiosis is, that isn't in dispute here. >College admissions departments already check to see if applicants are lying about race, e.g. by looking up their social media. Isn't that just checking for consistency? Say I change my mind about what race I am due to it being typically poorly defined. Many children have changing views about race over time, so it seems quite absurd to expect a literal child applying for college applications to fully understand and stay consistent about a topic that is constantly evolving and often nebulous. What you know at age 13 when opening your first facebook may not match your understanding of yourself at 17 when applying for college. And it's not clear to me what objective definition we are using of 'races' to make sure every applicant is following the same standard for determining when the applicant is lying and when they are not.
- SamReidHughes 5y ago
  Like, people's self-identified race is virtually always consistent with their ancestry, except when they're mistaken, unhinged, or taking advantage of lower standards for non-whites. 23andme's proprietary algorithm might be bad at predicting whether Elizabeth Warren calls herself native american or white, even if it it caught that 1/2^n wisp of DNA, but her self-identification is made on the basis of ancestry. > Say I change my mind about what race I am due to it being typically poorly defined. But it isn't poorly defined. Your race is the combination of your parents'. You're trying to use some college definition of race, and I don't know what sort of indoctrination made you that way, but that's not what people actually mean by race. Like, middle easterners who are Republican might call themselves white, while politically involved Democrats might call themselves "people of color". I know two siblings who probably split down this divide. But even if they do, they're actually the same race, because they have the same parents.
  - notch656a 5y ago

Mordisquitos 5y ago

I am a biologist and a bioinformatician, though not currently active in those fields. There is no such thing as an "objective definition of race", and claiming that so-called race can be determined from one's DNA is indeed pseudo-science.

SamReidHughes 5y ago
Then why can you tell what race people are by looking at them? I claimed you can predict 99.9999% of people's self-identified race from their DNA. Is that wrong? How high does that number have to be before you'd call it non-pseudo-science?
carrionpigeon 5y ago
> claiming that so-called race can be determined from one's DNA is indeed pseudo-science. That's taking it way too far. Genotyping populations identifies variation, and in cases where populations that have been reproductively isolated for tens of thousands of years (e.g. western Europeans and west/central Africans) are now in proximity, there will be substantial variation that is idiosyncratic to each population. In the US, reproductively isolated populations were suddenly placed in the same territory, and taboos on inter-racial mating persisted until very recently (50% toleration in the mid-1990s). One can accurately tell whether an American is phenotypically white versus black (or mixed) by DNA. That isn't to say that our ability to predict accurately black-vs-white will remain. The more race-mixing, the less ability to accurately categorize. We'll instead move to a situation similar to the one in which 'white' people find themselves today, where they are mixes of x percent of a variety of populations and not categorically 'Dutch,' 'Irish,' 'Finnish,' etc. Race is a sociological concept, yes, but it reflects reality in that people are visibly different. These differences are genetic in origin. It just so happens that in some cases, (e.g. two reproductively isolated and visibly distinct populations suddenly in contact with one another), these social categorizations will map to biological ones. In other cases where there is convergent evolution in terms of physical appearance, they won't.

notch656a 5y ago

Can you refer me to the 23andme algorithm that determines race with 99.9999% parity with self-reported race? Where do they report this outcome? I'm unable to determine the algorithm or find this claim by 23andme just from the reference to '23andme algorithm.' I've been totally unable to find such claims made by 23andme, suggesting they predict 'race' with any certainty to self-reported 'race.' I know what meiosis is, that isn't in dispute here. >College admissions departments already check to see if applicants are lying about race, e.g. by looking up their social media. Isn't that just checking for consistency? Say I change my mind about what race I am due to it being typically poorly defined. Many children have changing views about race over time, so it seems quite absurd to expect a literal child applying for college applications to fully understand and stay consistent about a topic that is constantly evolving and often nebulous. What you know at age 13 when opening your first facebook may not match your understanding of yourself at 17 when applying for college. And it's not clear to me what objective definition we are using of 'races' to make sure every applicant is following the same standard for determining when the applicant is lying and when they are not.

SamReidHughes 5y ago
Like, people's self-identified race is virtually always consistent with their ancestry, except when they're mistaken, unhinged, or taking advantage of lower standards for non-whites. 23andme's proprietary algorithm might be bad at predicting whether Elizabeth Warren calls herself native american or white, even if it it caught that 1/2^n wisp of DNA, but her self-identification is made on the basis of ancestry. > Say I change my mind about what race I am due to it being typically poorly defined. But it isn't poorly defined. Your race is the combination of your parents'. You're trying to use some college definition of race, and I don't know what sort of indoctrination made you that way, but that's not what people actually mean by race. Like, middle easterners who are Republican might call themselves white, while politically involved Democrats might call themselves "people of color". I know two siblings who probably split down this divide. But even if they do, they're actually the same race, because they have the same parents.
- notch656a 5y ago