Monthly Archives: March 2018

If you have Netflix I highly recommend Ugly Delicious. I watched it over the past week, as opposed to binging, and that probably made it easier, because the visuals and display of gustatory prowess can be a bit much. Many viewers will roll their eyes at the excessive seriousness and constant attempt to hook food […]

Read more

A long post at my other blog, The Maturation Of The South Asian Genetic Landscape, a reflection on the important preprint The Genomic Formation of South and Central Asia. Shorter: The original inhabitants of the Indian subcontinent who descent from the “out of Africa” migration separated very quickly, ~50,000 years ago, from other eastern populations … Continue reading “South Asian genetics, the penultimate chapter”

Read more

The above is a stylized map from the preprint, The Genomic Formation of South and Central Asia. In broad strokes, it says some things that are very expected, and some things that are not so expected. The abstract is long, but I’ll reproduce it in full: The genetic formation of Central and South Asian populations […]

Read more

I don’t know most of the people who contribute to this weblog anymore. So I don’t know how to contact you. Can you please update your profile with an image icon so that it’s easy to see who is who?
Thanks.

Read more

Since Who We Are and How We Got Here is out I thought I would spoil the “India chapter” (though you should buy the book!). – The “Ancestral North Indians” are best modeled as a 50/50 ratio of Yamna-type people from the steppes & “Iranian farmers.” The implication is that the Indo-Aryans mixed with agriculturalists … Continue reading “The Indian chapter of Who We Are and How We Got Here

Read more

A few years ago I watched a bunch of Megan Bowen’s YouTubes about living in Korea as an expat. In one episode she had explained that the reason she had a black American accent (she’s from Georgia I think) is that she is a black American. Just a very light-skinned one. In other videos, you […]

Read more

Yesterday on Twitter I made a quip about “linear Western models of time.” A friend pointed out that that was actually “Judeo-Christian.” I was going to agree…but then I realized something: I vaguely recalled that eschatology and millenarianism were things that some have hypothesized came into Judaism from Zoroastrianism. The historical context is straightforward. The […]

Read more

From the comments: Something that confused me very early on in the book- the San are shown branching off from the rest of humanity prior to Mitochondrial Eve. How can Eve be a common ancestor in this case? Admixture? The commenter is talking about an early portion of Who We Are and How We Got […]

Read more

This week Spencer and Razib talk to James Lu, co-founder of Helix, about the past, present and future of the “app store for personal genomics.”

Read more

The classical model of genetic inheritance, which dates back to the 19th century, involves discrete traits that are transmitted across generations in easily detectable patterns. These patterns follow what is known as a “Mendelian” model of inheritance: first adopted by the Austrian monk Gregor Mendel, based on his work with pea breeding in the mid-19th century. Largely unrecognized during his lifetime, Mendel’s work was rediscovered in the early 20th century, when biologists adopted it to explain patterns of genetic inheritance. These were the first professional geneticists.

Punnet square for pea pod color

The Mendelian framework is simple. Most complex organisms have two copies of a given gene. These are alleles. The alleles can come in particular varieties. In the example to the left, you see that pea pods come in two colors, and there are two alleles, yellow Y and green y. Because the expression of y is recessive, plants with green pea pods can only produce other plants with green pea pods. In contrast, since Y expresses dominantly, meaning you only need a single copy for the trait to be expressed, two plants with yellow pea pods can potentially give rise to both colors.

This elegant simplicity of Mendelian genetics did two things. First, it replaced the prescientific “blending” model of inheritance, where the physical characteristics of parents “mix” together to produce offspring. For example, they thought cross pollination of a red and a white flower would produce a pink one. This model, which makes intuitive sense, could never explain how populations retained all their variation — as opposed to being blended away. Second, the theoretical basis of modern population and medical genetics derives from Mendelian principles.

Malia, 6’1″ (left) is taller than her mother Michelle, 5’11”

Diseases like Cystic Fibrosis and Tay-Sachs express in individuals who carry two copies of the malfunctioning gene. They’re recessive Mendelian diseases, and diseases like these are the basis of much of modern medical genetics.

Not all traits and diseases, however, are like this. Neither height nor risk of developing schizophrenia have a simple inheritance pattern. You probably know tall offspring of short parents, and vice versa. Many people with schizophrenia do not have parents with schizophrenia, and many people who develop schizophrenia don’t have children who later develop it. You can’t just look at a pedigree and figure out simple probabilities, like you can with many simple traits and Mendelian diseases.

We also know that these characteristics are heritable within the population. That is, if you look at the characteristic in parents and offspring, they tend to be positively correlated. It’s not a perfect correlation, but the relationship is strong enough to suggest that there is a genetic disposition within individuals.

The heritability of height and schizophrenia is about 80% or more. That means 80% or more of the variation of the trait in the population is due to variation of genes in the population. This is a statistical inference.

How does this relate to Mendelism?

In the early 20th century, population geneticists realized that if you assume that there are alleles at many different genes, the combined action of those alleles could result in a distribution of outcomes. Instead of two flavors at one gene, there could be different variants across hundreds or thousands of genes. This results in an increase in a wide range of outcomes, as opposed to just two outcomes (as with recessive diseases) — potentially thousands of positions along a spectrum.

There aren’t two categories, tall or short, but a value of how tall you are. It’s a quantitative trait.

Correlation between height prediction and true height

Before modern genomics, which gave us access to the fine-grained variation across the DNA sequence, and computing, which allows for powerful analysis of large quantities of data, we couldn’t relate quantitative traits to individual genes. Luckily, we live in a world where modern genomics and computing do exist. Researchers can look at patterns of variation across the whole genome of many individuals in populations and how that relates to variation in characteristics. A new method can predict 40% of the variation in height simply from the genetic sequence!

Like height, schizophrenia is a highly heritable trait whose variation is controlled by many genes. But unlike height, for schizophrenia we are more interested in the likelihood of developing the disease rather than any quantitative value. Therefore, researchers will construct a “polygenic risk score,” which takes all your “risk” alleles across your genome, and combines them to produce and overall risk against the population norm.

Researchers do this by looking at associations between schizophrenia and particular markers. They assemble a list of markers that are associated with schizophrenia to a level of statistical significance. Then they look at another dataset, and test their markers to see if it predicts the variation of risk within that dataset across individuals. In this way researchers can construct a list of valid markers from the genome and use them to give individually tailored risks.

All of this is state-of-the-art and in its early phase. Most individuals haven’t been genotyped, much less had their genomes sequenced. You can’t construct a risk score on an individual who doesn’t have their genetic data. Additionally, many of these predictions are sensitive to the populations that they were tested in: there is still much medical research that needs to be done in people of non-European origin. Finally, unlike many Mendelian diseases, polygenic risk prediction may not be definitive enough for anyone to act on the results in any concrete manner. If you are told you have a two times greater risk of developing a disease where the average person has a 5% risk, does knowing you are at 10% change your behavior at all?

But knowing who is, and isn’t, as risk within the population is useful for public health and prevention, and probably saves money and lives if you add up all of the individuals. In the end, this is the “big data” direction our society is heading.

Learn more about where your traits for food tolerance fall on the spectrum and explore your Metabolism story today.

https://medium.com/media/c11cc60d68e5a729bd45957dd5bd31ec/href


Complex traits to individual predictions was originally published in Insitome on Medium, where people are continuing the conversation by highlighting and responding to this story.

Read more

New preprint, Something old, something borrowed: Admixture and adaptation in human evolution. This part jumped out at me: …Indeed, for most traits, the contribution of archaic human alleles to present-day human phenotypic variation is not significantly larger than those of randomly drawn non-introgressed alleles occurring at the same frequency in modern humans. Interestingly, in both […]

Read more

For a while now I’ve been telling my cofounders at DNAGeeks.com that we need to do a shirt about reading supplements. If you don’t know what I’m talking about, you’re a heathen. Basically, if I blog about a paper, I try and read the supplements, because there’s a lot of good stuff (often the method […]

Read more

Does anyone have a galley or review copy of She Has Her Mother’s Laugh: The Powers, Perversions, and Potential of Heredity? The book is long, and review copies are in short supply. Would be nice if I could see it before it’s released at the end of May. Just use my contact email if you […]

Read more

Reality, it turns out, is more complex and interesting than scientists ever imagined.

Read more

I read Benedict Anderson’s Imagined Communities: Reflections on the origins and spread of Nationalism because other people read it. This is a book that is routinely alluded to in discussions by pundits of various stripes. On the back of the 2006 edition, the publisher notes that over 250,000 copies have been printed of this short academic […]

Read more

Last year’s The time and place of European admixture in Ashkenazi Jewish history is very close to the last word on the genetics of the ethnogenesis of Ashkenazi Jews. Here’s the author summary: The Ashkenazi Jewish population has resided in Europe for much of its 1000-year existence. However, its ethnic and geographic origins are controversial, […]

Read more

A few people have been pointing me to a new paper, A Bayesian phylogenetic study of the Dravidian language family, which implies that the Dravidian language family diversified ~4,500 years ago. I don’t have much to say about the paper itself since it aligns with my own conclusions, but it’s well outside of any field […]

Read more

Back when I was in college one of my roommates was taking a Chinese philosophy class for a general education requirement. A double major in mathematics and economics (he went on to get an economics Ph.D.) he found the lack of formal rigor in the field rather maddening. I thought this was fair, but I […]

Read more

I haven’t posted on one of these in a while. Mostly because I don’t know what to say about Henan. Henan is where China began. As noted in Wikipedia four of the eight ancient capitals of China are located in this province, in the heart of the North China plain. Chineseness, as we understand it, […]

Read more

20/42
Razib Khan