News for hackers. Stuff that matters.
Re:Vitamin D uncorrelated with Covid severity in those of Eu... by Blikkentrekker on Wednesday January 20, 2021 @07:47PM (#25853919)
What exactly am I supposed to read into “European descent”.

For something so prominently featured in the title, implying some significance to it, I cannot find any further explanation of it:

> We predominately used previously published and freely available data for this study. The study by Jiang et al 19 is a meta-analysis of GWAs of vitamin D levels carried out using participants of European descent. The COVID-19 Host Genetics Initiative20 uses data from multiple cohort studies,21–26 including the UK Biobank (UKB).

Does it mean “U.K. residents”, “European residents”, or “persons whose ancestors had European residence, wherever they be located now.” and to how many generations back is this?

Re:Vitamin D uncorrelated with Covid severity in those of Eu... by throwaway2245 on Wednesday January 20, 2021 @07:55PM (#25853998)
It is a genetics term of art that means "white people".

This is particularly important for Vitamin D, since the colour of your skin has a meaningful influence on Vitamin D production.

The research article says "Our results also cannot be used to comment on the relationship between vitamin D and COVID-19 in non-Europeans." - they take no stance for non-Europeans.

This reflects the common point that most medical/genetic data is sourced from Europe and North America: there often isn't enough data for People of Colour to accurately comment on points of racial difference, whether white people have been sourced deliberately or inadvertently.

Re:Vitamin D uncorrelated with Covid severity in those of Eu... by Blikkentrekker on Wednesday January 20, 2021 @08:07PM (#25854091)
Methinks that would be letting scientific thought be influenced by cultural memes.

“race” is not really a distinction that has any actual biological merit and is purely a cultural creation, There is really no expectation that different “races” would be statistically differently affected any more than something other completely arbitrary and cultural such as “office workers" or “virgins”.

> This is particularly important for Vitamin D, since the colour of your skin has a meaningful influence on Vitamin D production.

Yes, but the lines between races is arbitrary and culturally drafted.

The same man who would be considered “black” in the U.K. would be considered “white” in Kenya. What test is even used as discriminant of “white people”? self-report, a pencil test?

In particular, I've seen many persons who are apparently considered “not white” in the U.S.A. whose skin complexion woud surely not have such a meaningful influence. Consider also the story with Rachel Dolezal, which showed well that race is a cultural meme, not a phænotypically defined set of traits, nor something upon which a biological significance can be built.

Re:Vitamin D uncorrelated with Covid severity in those of Eu... by throwaway2245 on Wednesday January 20, 2021 @08:38PM (#25854279)
I can put this another way which avoids race politics:

Genetic sample data is not truly randomised, and tends to greatly over-represent a particular subgroup of 10% of people who share common genetic markers (including recognisable phenotypic consequences such as: in general having white skin, and in general adults being tolerant to lactose) that are not globally common. For convenience, we'll say that groups of people with those markers are "of European descent".

The extent of the over-representation is so great that edge cases are likely to have no impact in changing the overall bias of the data.

As such, conclusions drawn globally should be treated at least with caution - and with suspicion when one of those genetic markers is highly relevant to the conclusion.

Re:Vitamin D uncorrelated with Covid severity in those of Eu... by Blikkentrekker on Thursday January 21, 2021 @06:20AM (#25857758)
Then they should simply word how they drew their sampling such as “U.K. residents” or however else they drew it.

I'm still not sure as to what exactly the sampling is here and what “European descent” communicates? The way I see it it can mean any of the following:

- Participants were asked to self-report their “race”, any that did not report “white” were excluded.

- Participants were given a pencil test to determine their “whiteness”, those failing were excluded.

- Participation was drawn from U.K. residents

> The extent of the over-representation is so great that edge cases are likely to have no impact in changing the overall bias of the data.

You say so, but I feel this distinction is more a cultural meme than biologically meritful. I for instance rarely see, prominently in the title “in those of Boston descent” when all participants were selected from Boston, which could produce similar biases.

I feel it is likely that it is highlighted because they feel that “race” is more meaningful than it biologically is, and are attempting to draw inferences from it.

Re:Vitamin D uncorrelated with Covid severity in those of Eu... by brigandish on Thursday January 21, 2021 @08:57AM (#25858793)
> I feel it is likely that it is highlighted because they feel that “race” is more meaningful than it biologically is, and are attempting to draw inferences from it.

Try telling that to sufferers of sickle cell anaemia.

They’re not making a general statement about race, they’re relating it to a specific thing. “Race” may be an overloaded, and often misused term - certainly a misnomer - but it still can be used as a shorthand that is useful especially in medical contexts. Here, it’s probably to warn against generalising beyond whites, though the meaningfulness will be borne out by research with regards to this specific virus, not through your feelings or inferences about possible inferences.

Re:Vitamin D uncorrelated with Covid severity in those of Eu... by Blikkentrekker on Thursday January 21, 2021 @09:47AM (#25859339)
> Try telling that to sufferers of sickle cell anaemia.

And do you believe that the persons who in various cultures would be called “black” but are actually genetically 95+% “white” would be more likely to suffer from that? or conversely that the persons who are genetically 95% “black” but would in various cultures be called “white” are not?

Such risk factors do not of course care about arbitrary cultural racial classifications which are often noted to be asymmetrically contagious in various cultures.

> They’re not making a general statement about race, they’re relating it to a specific thing. “Race” may be an overloaded, and often misused term - certainly a misnomer - but it still can be used as a shorthand that is useful especially in medical contexts. Here, it’s probably to warn against generalising beyond whites, though the meaningfulness will be borne out by research with regards to this specific virus, not through your feelings or inferences about possible inferences.

If that be their intent, they firstly communicate it poorly and should have simply said so, and secondly it's useless.

Again, in various countries with a majority nonwhite population, a person who might genetically be 5% white, and 95% local population could conceivably be termed “white”: — is it their intention and your contention that the results of this research may then be applied to such a person that is effectively genetically close to local population?

Re:Vitamin D uncorrelated with Covid severity in those of Eu... by brigandish on Thursday January 21, 2021 @07:16PM (#25866585)
> And do you believe that the persons who in various cultures would be called “black” but are actually genetically 95+% “white” would be more likely to suffer from that?

Individuals and populations are different.

> Such risk factors do not of course care about arbitrary cultural racial classifications which are often noted to be asymmetrically contagious in various cultures.

Risk factors may not care but individuals in at-risk populations are likely care that knowledge of different risks to them, based on their race or not, is available.

> If that be their intent, they firstly communicate it poorly and should have simply said so, and secondly it's useless.

That's your contention, I'm yet to see a cogent argument or any evidence for this.

> Again, in various countries with a majority nonwhite population, a person who might genetically be 5% white, and 95% local population could conceivably be termed “white”: — is it their intention and your contention that the results of this research may then be applied to such a person that is effectively genetically close to local population?

Reading this, it is my belief that you have missed the point of the authors intent, the reasons for using race as a category in medicine, and all of the arguments put forth to you.

Re:Vitamin D uncorrelated with Covid severity in those of Eu... by throwaway2245 on Thursday January 21, 2021 @09:07AM (#25858906)
> Then they should simply word how they drew their sampling

They do literally cite in words where they got their sample data in the first two sentences of the linked article.

This, and referencing an apartheid test for ethnicity, suggests you are no longer speaking in good faith. I have already explained why this is biologically highly relevant to this case.

There is no particular additional harm in claiming this research result applies in general to people with European descent, a known problem of bias in existing data sources that is known to have an impact on the quality of medical care, and whereby you can make your own decision as to whether this might apply globally.

There is potential additional harm in claiming this research result applies globally, since you are working with highly biased data, and applying solutions that appear to work for a distinct subset of the population might work ok but also might be harmful to the other 90% of people.

Re:Vitamin D uncorrelated with Covid severity in those of Eu... by Blikkentrekker on Thursday January 21, 2021 @09:42AM (#25859287)
> They do literally cite in words where they got their sample data in the first two sentences of the linked article.

No, they do not. They say they used the U.K. biobank, but otherwise do not tell how they selected upon “European ancestry" in that bank.

> This, and referencing an apartheid test for ethnicity, suggests you are no longer speaking in good faith. I have already explained why this is biologically highly relevant to this case.

And I have argued why it isn't by the argument that an individual that is biologically “white” or almost fully white can culturally be considered “black”, depending on the culture, and obviously vitamin-D production cannot be assumed to care about such cultural divisions.

It has long been noted that many persons in the Anglo-Saxon world, particularly the U.S.A. that are referred to as “black” are genetically speaking effectively “white” and almost completely “of European ancestry”; — the article and those who replied to me are entirely unclear whether they are included or not in the sample.

> There is no harm in claiming this research result applies in general to people with European descent, a known problem of bias in existing data sources that is known to have an impact on the quality of medical care, and whereby you can make your own decision as to whether this might apply globally.

There is absolute harm in using vague terms of which no one exactly knows what it means.

What do you, for instance believe the actual selection of “with European ancestry" signifies here? What concrete, testable definition of “of European ancestry” was used?

> There is potential harm in claiming this research result applies globally, since you are working with highly biased data, and applying solutions that appear to work for a distinct subset of the population might work ok but also might be harmful to the other 90% of people.

Indeed there is, but there is no false dilemma here. — both can be as ridiculous as the other and in fact do not exclude each other, because the term “ancestry” is used, not “residence” which indeed implies that the result holds for persons of European descent, globally. One can be located anywhere in the world and still be “with European ancestry”.

Re:Vitamin D uncorrelated with Covid severity in those of Eu... by throwaway2245 on Thursday January 21, 2021 @10:03AM (#25859509)
> They say they used the U.K. biobank, but otherwise do not tell how they selected upon “European ancestry" in that bank.

The UK Biobank study is linked (and also has its own wikipedia page in any case).

This is simply how science research works. They are not repeating the information given in the source.

Re:Vitamin D uncorrelated with Covid severity in those of Eu... by Blikkentrekker on Thursday January 21, 2021 @10:16AM (#25859660)
And if they are to be found there what criteria it, if any uses to determine “with European ancestry” then I'm sure you can produce it, because I searched and couldn't.

And I find that in practice if research include a vague criterion in it's very title, it tends to come with far clearer definitions of that criterion in it's body than this.

Re:Vitamin D uncorrelated with Covid severity in those of Eu... by throwaway2245 on Thursday January 21, 2021 @11:00AM (#25860238)
I did find it for UK Biobank (one of several sources) earlier, quite easily, by clicking on the relevant link. I suspect that you could not find it (and did not know) only because you don't have relevant training/expertise in science/genetic research - in line with your first question in this thread.

They asked people in the UK in a digitally supported questionnaire/interview, for their ethnicity.

(Wikipedia also mentions that UK Biobank is criticised for sampling bias towards wealthy white Brits)

Re:Vitamin D uncorrelated with Covid severity in those of Eu... by calibas on Wednesday January 20, 2021 @08:54PM (#25854388)
It's fundamentally problematic, but some distinction needs to be made based upon genetics. Otherwise we do things like making dairy it's own "food group" and treating lactose intolerance like a disease.
Re:Vitamin D uncorrelated with Covid severity in those of Eu... by Blikkentrekker on Thursday January 21, 2021 @06:22AM (#25857763)
And was this distinction made upon genetics? were genetic tests performed.

As I said in another comment: I am entirely unsure how “of European descent” was selected upon and the research does not make it clear at all what it means by it, and it could mean any number of things.

For such a vague condition to be so prominently featured in the title, they could at least define how they selected upon it, if at all.