Zip code better predictor of health than genetic code | Hacker News Reader

Zip code better predictor of health than genetic code

hsph.harvard.edu

48 pointsmnkc16 comments

What a massive claim! Of course, after carefully reading the article several times just to make sure, it's just one attention-grabbing line said in one speech, by one person.

To prove this sort of claim, you'd need to challenge the millions of hours of research geneticists have done, all the open data they've published, all the medical advances they've fueled. Instead, without even considering that the claim might be wrong, without any scientific humility, without linking to any supporting research, Harvard confidently makes the most extreme claim; that zip codes have more predictive power than genes!

Disgraceful and absurd that Harvard would publish something like this. Disgraceful that I believed it even for a moment.

This person specializes in biostatistics: http://www.rwjf-newconnections.org/funded-scholars/melody-s-... Based on the list of her publications, I doubt she has extensive knowledge of genetics. Nevertheless, if she stopped for a moment, she'd have realized her statement might make no sense whatsoever. This title and statement should be flagged for inaccuracy.

To be fair, the predictive power of many different polygenic risk models is very low. (Source: doing a PhD in Genetics atm.) So this does not really surprise me. Usually about 50% of variance can be explained by genetics, and 50% by environment. Meaning that if you have a good enviromental predictor, such as a zip code, you may be able to predict a trait better, than with genetics, which contain a lot of cryptic relationships that a predictor may not integrate. Its totally fair to be critical of the merit of genetic predictions. Usually they are super bad.

While it's true this article is a whole lot of nothing, here's a better article with concrete facts on the same subject: https://www.scientificamerican.com/article/u-s-life-expectan...

On its face, it doesn't seem that crazy. Location is linked to both wealth and environment, which are both pretty significant factors in health. Do you think most doctors would say that average genetic variation makes a bigger difference than, say, poor childhood nutrition, or smoking?

Note that the title does not state that genetic code does not predict anything useful. Rather that zipcode is a significantly bigger factor, dwarfing variance seen by genetic code alone.

I have no idea if that is true, but that is one way of reading the title without disregarding all the work done with genetics.

At present, the claim made in the article is not as controversial within public health as it may seem: zip codes are a better predictor of many health outcomes, including life expectancy, than our present genome-based predictive models. The public tends to over-estimate the predictive value of polygenic risk scores, especially for complex conditions like metabolic syndrome. Zip codes capture a great deal of environmental variance, plus a bit of genetic clustering by shared ancestry. Most polygenic models fail to capture all of the relevant genetic variance, and are uninformative about environmental factors. Additionally, for some conditions, the magnitude of environmental variance and relative importance of environmental factors is greater than the variance and effect size of genetic factors.

The original article did not provide any supporting evidence for this claim. Can you point us toward actual research to back up your statement?

There are very few genetic risk score models that outperform traditional observational disease markers (here[1] is a non-paywalled discussion of cardiovascular GRS performance as an example). The best GRS results tend to be in relatively genetically-homogeneous populations that are similar to the population in which the GRS model was developed. In some cases, knowing the ethnicity or simple family history of a patient can buy you a good portion of the AUC of the relevant GRS.

So if you have a classifier like ZIP, that (1) epidemiologists have done a bit of legwork correlating to classical markers like obesity (or their correlates, such as income and dietary/smoking/prescription patterns) and (2) tends to follow familial/ethnicity clusters in (3) a heterogeneous population, you can amass a fair bit of predictive power on the cheap for complex disorders where environmental variance plays a role, as well as beating the spread on behaviourally-determined mortality/morbidity factors.

It is likely that the predictive power of GRS-based approaches will improve for many conditions in the future (they are of course already powerful for Mendelian disorders).

[1]https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4527979/

I don't have the slightest doubt that a classifier like zip code (or traditional observational disease markers) delivers much greater value than a genetic risk score model because its accuracy is nearly as good while its cost is a couple of orders of magnitude less.

I do have some doubts that the zip code actually gives BETTER prediction than the genetic risk score. I have difficulty believing that if someone had done a genetic profile on a patient and was willing to tell me either the patient's zip code OR the genetic risk score, that I would be better off asking for the zip code because it had greater predictive value. It' certainly not impossible (because of environmental factors that correlate with zip code), but it is surprising and I haven't yet seen actual research supporting it.

Why do you have those doubts? Do you have any "actual research" to support them?

As someone who doesn't really care either way, I will say that his arguments have been more convincing than you simply casting doubt and demanding proof based on what seems to be a gut feeling.

Bear in mind that people tend to live near family and near people of the same ethnic group as them. So not only does zipcodes include info on how rich you are, and therefore your access to healthcare and good nutrition, it also contains some information about your genetics.

The linked article just quotes someone saying this with no citation.

Probabilistic models based on naive gross-over-simplification of wastly complex reality at its best.

There is indeed a huge distance between DNA and non-genetic diseases and attempt to bridge this gap with a naive probabilistic model will surely yield nonsense.

Social and environmental factors are much more fundamental than actual DNA (gene regulation is still poorly understood) for non-genetic diseases. So, yes, statistically it is true.

This may be quite useful to health insurance actuaries in the years to come, depending on state laws. Though I imagine in the long run, they would vastly prefer a fully sequenced genome for all insureds, or IQ tests.

I wonder if we'll see those kinds of things in the future. I guess on a long enough timeline, the question isn't if, but when.

isn't sequencing genomes expensive though? what if they could have 90% of predictive power for the price of a filled out zipcode instead of requiring everybody to do lab tests?

It is expensive currently. It's dramatically decreased in price over time, though. ($2.3B in 2003, to $1000 in 2014[0])

[0] https://techcrunch.com/2017/01/10/illumina-wants-to-sequence...

Reply on news.ycombinator.com