December 30, 2010
More Genes For Height Identified

64 height-associated variants identified.

To search for genetic variants associated with adult height, researchers performed a complex genetic analysis of more than 100,000 individuals. "We set out to replicate previous genetic associations with height and to find relevant genomic locations not previously thought to underpin this complex trait" explains Dr. Brendan Keating, also from The Children's Hospital of Philadelphia. The authors report that they identified 64 height-associated variants, two of which would not have been observed without such a large sample size and the inclusion of direct genotyping of uncommon single-nucleotide polymorphisms (SNPs). A SNP is a variation in just one nucleotide of a genetic sequence; think of it as a spelling change affecting just one letter in an uncommonly long word.

Note the number of height-associated variants. The reason it has taken so long to make use of genetic sequencing data is that many different variants each contribute to height, personality, disease incidence, and other attributes and risks. Since each variant contributes so little a large number of subjects must be used to detect the various small contributors.

I want the search for the meaning of genetic variants to go much faster. With prices for personal genetic testing getting pretty cheap I'm intrigued by the prospect that lots of individuals can pay affordable prices to get themselves tested and also enter health info onto a web form to participate in scientific research into the meaning of genetic variations. This holds the promise of greatly lowering the cost of research while speeding up the rate of discoveries.

The first phase of the study includes development and validation of web-based surveys to assess the drug side effects and drug effectiveness experienced directly by 23andMe's customers. During the second phase, the research team will determine whether this approach enables them to replicate previously known associations between response to these three classes of drugs and variation within two genes: CYP2C9 and CYP2C19. 23andMe's research team will also search for previously unknown genetic factors associated with response to these classes of drugs, taking into consideration a broad range of non-genetic factors such as age, sex, and body-mass index, among others.

In previous studies, 23andMe has demonstrated that self-reported information from customers yields data of quality comparable to that gathered using traditional research methods.

23andMe's new gene chip tests about 1 million locations on genes where variations occur. Suppose a few hundred thousand people sign up for their service and then also enter in lots of information (e.g. height, eye color, hair color, weight, assorted other physical measurements, history of allergies, asthma, injuries, drug reactions, and other events and maladies). The influences of large numbers of genetic variants could be discovered.

Lots of discoveries are waiting to be made using cheap gene chips to collect data on lots of known single letter genetic variants. After that flood of genetic data comes an even bigger flood. While sequencing the first complete genome took years and hundreds of millions of dollars the costs have now fallen into the tens of thousands (if not lower) per genome. With these low costs the rate of full genome sequencing has now reached hundreds per quarter and therefore over one thousand per year. We can expect the rate to go up by more orders of magnitude.

Complete Genomics said last week that it anticipates sequencing more than 300 human genomes in the fourth quarter of 2010.

Full genome sequencing allows identification of extremely rare genetic variants and also the measurement of large copy variations and other genetic variants not easily captured by gene chips. But data the gene chips alone, used by hundreds of thousands and even millions of people and combined with personal data, will allow scientists to identify at least tens of thousands of genetic variants that influence who we are.

Share |      Randall Parker, 2010 December 30 10:45 PM  Human Population Genetics

Someguy said at December 31, 2010 6:44 AM:

This is far from my knowledge base, but wouldn't it make sense to try and identify genes for the length of specific body sections. Height is a composite of the length of a number of sections of the body. I have a long torso and neck, but my legs are of average length. Do genes impact overall height or simply the length of specific bones?

bbartlog said at December 31, 2010 11:27 AM:

'Do genes impact overall height or simply the length of specific bones?'

Logic dictates that there must be genes of both classes.

Aki_Izayoi said at January 1, 2011 11:51 PM:

Eye color and hair color are not necessary to enter because the origin of its loci responsible for the majority of the variation isn't much of an enigma:

"For example, consider the HERC2-OCA2 region of the genome in West Eurasians. It is responsibe for ~75% of the variation in blue vs. non-blue eye color. It also seems to be responsible for some variation in skin complexion. This makes sense, as one would presume that different pigmentation related phenotypes would be controlled by the same set of genes. Finally, the genomic region is characterized by a very long haplotype, which might be an indication of rather recent natural selection driving one variant to high frequency. More plainly, it seems probable that the genetic variants around HERC2-OCA2 which are correlated with light eye colors rose to high frequency with the last 10,000 years or so."

As for height, I suppose one should just look for loci that contribute to overall height as opposed to the length of specific bones (which contribute to height). Likewise, Robert Plomin decided that it is more productive to search for the loci responsible for the variance of g in the general population rather than specific aptitudes.

Randall Parker said at January 2, 2011 9:15 AM:


You are back!

Good post by Razib.

To your point: There's still the minority of that variation. Plus, are the alleles for other eye color variants known? What causes, for example, a greenish hue with a gold inner circle? (that'd be me)

This raises a bigger point: A web-based system to collect physical measurements of people would need integration with cameras and other sensing devices. For example, uploaded eye pictures could get processed to extract out a lot more details, not just shade of black or blue but also patterns in the streaking and mixtures of colors.

A picture of a hand could be processed to get the ratio between finger lengths. Pictures of the face from different angles could measure nose proportions, size of chin (a proxy for masculinity/feminity as is index/ring figure ratio), and other features.

I'd like to see a Human Measurement Project that has a big web infrastructure for collecting lots of details of people and allowing them to be tested in a variety of ways. Then compare all that information against genetic testing results and see what turns up.

Post a comment
Name (not anon or anonymous):
Email Address:
Remember info?

Go Read More Posts On FuturePundit
Site Traffic Info
The contents of this site are copyright