deCODE Publishes Largest Human Genome Population Study

March 25, 2015

By Allison Proffitt

March 25, 2015 | With a package of four papers published today in Nature Genetics, deCODE genetics presents the largest set of  human genomes from one population, and findings including a set of complete human knockouts, a new timeline for the common Y-chromosome ancestor, and loss-of-function variants that confer Alzheimer’s disease risk.

The project reflects a population genetics approach that deCODE founder and CEO Kari Stefansson believes should serve as a model for other large-scale population studies including President Obama’s Precision Medicine Initiative.

deCODE conducted whole genome sequencing on 2,636 individuals—a little less than 1% of Iceland’s population. At that rate, any sequenced individual is 6-8 generations from anyone else, Stefansson said, allowing the deCODE team to impute sequence variance down to .1%. If related individuals are sharing common variants, then we only need to sequence particular regions in one individual to infer that others carry the same variant, he said, adding that that assumption was validated by further sequencing of a total of 104,000 people.

Iceland’s population is uniquely suited to this type of study. The population is fairly small though not inbred and extremely well-documented. Most Icelanders have known family pedigrees going back many generations.

Though deCODE has a history of developing robust and innovative tools for population genomics—deCODE’s informatics platform is now available from an independent company, WuXi NextCODE Genomics—the country’s population itself has always been one of deCODE’s biggest advantages. Stefansson and his team are making the most of it.

One of the most valuable datasets to emerge from the work is a large set of complete human knockouts—individuals who carry a loss-of-function mutation in both alleles. deCODE researchers found that between 7.7% to 8% of the Icelandic population has at least one gene knocked out, a rate Stefansson believes will be at least as high in other populations.

“I am absolutely convinced that when people start to look at the same in more outbred populations—bigger populations—the number of genes where we are going to see knockouts is going to grow,” he said.

As mouse knockouts are important models for research, human knockouts will also prove to be very valuable to research, especially when compared to phenotypic data, to identify biochemical pathways.

Such a high rate of human knockouts may be surprising. With that many loss-of-function mutations in a population, how do we survive?

“This is constantly happening,” Stefansson said. “Our data is constantly humiliating us.”

But Stefansson believes that further research will certainly reveal protective loss-of-function mutations and biochemical redundancies, or “safety nets”, in addition to deleterious mutations.

Thus the more data, the better. Stefansson is quick to point out that even this effort, which he calls a “fairly detailed insight into the sequence of the genome of an entire generation” is far from complete. Average read length in the sequencing was 125 nucleotides, which Stefansson admits is not likely capturing structural variations. Stefansson said his team does Sanger sequencing on regions of poor coverage.

A Model Partnership

The value of the work goes far beyond just Iceland, Stefansson said. “This is very much more than a molecular national selfie.” In addition to data about disease genes, the work will help develop new understanding of the history of populations, and human history in general, he said.

One of the four papers presented loss-of-function variants in ABCA7 that confer additional risk of Alzheimer’s disease, an area of research deCODE has been engaged in for years.

Another reported new findings for the Y-chromosome mutation rate. Previously, the common ancestor of all Y chromosomes was believed to be about 338,000 years old but deCODE’s data suggests that the common male ancestor was living about 239,000 years ago. The new date more closely aligns with the current date of a mitochondrial ancestor.

Stefansson joked, “It was really nice for us, because we males looked in the mirror and we simply knew we were not this old.”

In fact, Stefansson was universally jovial in a press briefing earlier this week. When asked by a reporter why some of the Alzheimer’s findings had not been discovered before, Stefansson answered: “Mostly because we are the best, and we find these things.”

According to Stefansson deCODE’s expertise has been vitally important to Amgen since the pharma’s acquisition of the Icelandic firm in late 2012. Drug discovery programs have been both canceled and launched at Amgen thanks to deCODE’s analysis and data validation. In fact, deCODE is having such a major impact on drug discovery at Amgen, that other pharmas are working to replicate the arrangement, Stefansson said.

It’s a glowing assessment of long transition. deCODE filed bankruptcy in 2008, was acquired by Amgen for $415 million in 2012, and spun out its comprehensive clinical data IT infrastructure and interpretation suite as NextCODE Health in 2013. Early this year, NextCODE was bought by WuXi PharmaTech and rebranded WuXi NextCODE Genomics.

The original goal for NextCODE was to bring the data infrastructure and software tools developed at deCODE to the broader diagnostics and big genomics community, and the company counts the most recent research as a win for the platforms.

“The sequencing revolution promises to transform the power and practice of medicine, and deCODE’s unique assets and approach in Iceland are giving the world a picture of just how much this revolution can deliver,” said Jeffrey Gulcher, co-founder and Chief Scientific Officer of Wuxi NextCODE. “Our technology, which was originally developed at deCODE, is helping to do that by providing other groups in other populations with the same proven system whose power is on full display today.”

For his part, Stefansson is only encouraging of other populations that choose to take on the breadth of work deCODE has.

“Other countries are now preparing to undertake their own large-scale sequencing projects, and I would tell them the rewards are great,” he said.