If you’re a scientist at Berkeley, you’ve found a great place to be. There is no shortage of conferences, symposiums, lectures, and workshops in which to learn more about your field of interest. Today, the Center for Computational Biology hosts a symposium on Human Genome Variation, the first event to be held in the not-yet-re-opened Li Ka Shing Center for Biomedical Health. Stay tuned as I attempt my first live blog from this sold out symposium.
Rasmus Nielsen introduced this symposium as an opportunity to bring the best researchers in the fields of computational biology and population genetics together at Berkeley so that we can learn more about Human Genome Variation.
Homozygosity, Linkage Disequilibrium, and Imputation in Diverse Worldwide Populations: The first speaker, Noah Rosenberg, works on such problems as the structure of gene trees and what we can learn from these. Rosenberg talks about how we observe patterns of genetic variations within populations, and then try to understand these patterns using mathematics and population genetic theories to make assumptions. We can then use statistical tools and inference to learn something about human evolution and evolutionary processes.
Rosenberg gives some history: The first observations of genetic variation within human populations were made by the Hirschfields, who observed ABO blood variation in soldiers. Next we began to look at immunological polymorphisms. We then looked extensively at mitochondria, but this has its limitations. And then came the explosion of technology that allowed us to obtain exponential amounts of sequencing data and the computational advancements, that leads us to where we are today.
Here are some interesting points in Rosenberg’s talk:
Pairs of individuals from different populations are only slightly more genetically different than pairs from the same populations.
Geographic barriers (such as the Himalayas) explain continental clustering of genomes.
Spatial representations of SNP (single-nucleotide polymorphism) variation matches geography.
No one had looked into the origins of the 18 stem cell lines that were approved for use in federally founded studies for disease mechanisms, drug discovery, and cellular therapies. (Thanks, Bush, for placing a cap on scientific progression). Unsurprisingly, when their ancestry was analyzed, it was found that their diversity was very limited. The most commonly used lines were from European ancestry, and many of them came from the same mating pair, making these lines equivalent to siblings. The first wave of disease research was primarily based on European populations. Starting in 2008, we began to broaden our population representation.
Genetic diversity declines with distance from the source. Genetic variation in Africa is most diverse, indicating that the expansion of human populations originated somewhere in eastern or southern Africa.
His concluding remarks: There are genuine differences in disease risks due to biological variation. As diseases within a population persist, it will be important to understand genetic variation pattern associations with these diseases. We must acknowledge that our understanding of human evolution and variation plays a significant role in our understanding of human disease and our ability to prevent and combat these diseases.
Next up: Elaine Ostrander, who will talk about The Shape of Things: Genetics of Complex Traits in the Dog. I love all the different breeds of dogs – there is so much variation in man’s best friend. I’m excited to learn more. Ostrander has worked with dog genetics for over 20 years, so she’s quite the expert.
Ostrander reminds us that the closest ancestor to the domestic dog is the gray wolf. All breeds of dogs are members of the same species, despite their extreme morphological variation. Kennel clubs mandate genetic isolation of breeds, causing population bottlenecks; what a great model for studying isolated populations!
First, Ostrander started looking for genes that determined body size – she shows a picture of a Great Dane next to a Chihuahua – what a difference! Her lab ran a genome-wide scan in the Portuguese Water Dog, and used very precise quantitative data using skeletal metrics to quantify phenotype to link with genotype. They found that the gene IGF1 is correlated to size – i.e., breeders select for IGF1 mutants to obtain small breeds.
Alleles are fixed in breeds, so this mutation was fixed early on. They genotyped 374 grey wolves from 17 populations and 115 individuals from five distantly related wild dogs – and could not find this SNP. They then found Grey wolf haplotypes from Israel associated with small dogs. The small dog and all Israeli grey wolf haplotypes were ancestral to the large domestic dog haplotypes and all other grey wolf haplotypes, indicating ancestry in the middle east.
How do genotypes associated with size all fit together? There is a continuum of dog size across the species. Breeds between 5-15 lbs are homozygous for small alleles. Breeds between 16-30 lbs have a large allele for stc2, but small alleles for remaining QTLs. 31-50 lbs: large allele for all QTLs except IGF1. 50+ lbs, large allele in most dogs for all QTLs. There remains a large amount of size variation in mixed groups.
Moving on to proportions: Chondrodysplasia is a breed-defining trait for some domestic breeds (think Corgi, dachshund, and Basset hound). They found a retro-copy of the fibroblast growth factor 4 gene (fgf4), which was the first functional polymorphic retrogene segregating in a mammalian species to be found.
Next: Fur types, which can be silky, curly, course, short, long. They found that long versus short fur was caused by an alteration of a highly conserved Cys95Phe, and that curly versus straight is caused by a non-synonymous change.
How do all these variations interact to create our wildly diverse breeds? Ostrander shows a table comparing genotype with phenotype, complete with pictures of these different breeds (and lots of laughter). I disagree with her assertion of which breed is adorable (Bichon Frise).
And then she talks about shedding – and just like some men, there are some breeds that shed just along the top line. (Alas, my dog sheds EVERYWHERE).
She now talks about quantifying the canine skull and shows pictures of the English Bull Terrier, Pug, Cane Corso, and Afghan Hound. Wow, they all look very different! Her new favorite breed is the Tibetan Mastiff, which is the size of a small lion – they look just like dog-lions in the picture she shows. How to bring together size with shape? They collected data on 49 different landmarks measured with the microscribe, including angles, arches, rostrums and snouts. They have data from 750 skulls and counting. Surprisingly, many people have skull collections in their basements. They discovered that 67% of purebred skull variance is caused by PC1.
And talk about our accumulation of genome sequence, we now have whole genome sequencing of 12 phenotypically diverse dog breeds.
She now talks a little about developmental biology and ventral neural crest derivatives sensitized to Bmp3 knockdown, her labs newest project. They took this back to humans, and compared these findings to patients with microdeletions of BMP3.
In sum: There is strong selection for body size genes; IGF1 is the strongest contributor. Asymmetrical dwarfism is controlled by a single ancient and dog specific mutation. Combinations of alleles at three genes: pattern, length, and curl, account for 90% of fur phenotypes in dogs.
Next up is Carlos D. Bustamante, who will be talking about Population Genetics in the Personal Genome Era. This talk will be really interesting!
He starts out talking about the biases in our accumulation of information in human genome data: 96% of participants in large scale genomic studies are of European descent. How do we bring in other populations? How relevant is our knowledge of disease in one population to another? Example: HDL-cholesterol levels.
He first looks at Latin American diversity, the very interesting Taino community, and the distinction between population genetics and cultural definitions. According to Bustamante, race is not a biological concept – it is a sociological one, and an inappropriate model in describing human genetic populations. Goal of Taino Genome Project: assign continental origin for each contiguous genomic segment for 35 admixed Puerto Ricans.
1000 Genomes Pilot projects (TGP). The inferred joint, rare variant spectrum from TGP contains mostly variants private to a population. Whose rare variants are we going to quantify? I believe that the TGP point deserves its own blog post.
And for all of you wanting to know your own genome, this year we’ll be able to sequence your individual genome in under 2 hours for less than $1,000 bucks.
One last interesting study: blond hair in the Solomon Islands. Single gene: TYRP1.
And that concludes the first half of the Human Genome Variation 2012. The symposium continues for the remainder of the day, but I best get back to making my own scientific breakthroughs.
Individual lectures from this conference will be available online.