This is Part 2 of a two-part series. See Part 1.
November 19, 2013. The second half of the final plenary panel discussion on October 21 focused on the next steps needed for biological follow-up and clinical translation of the existing genetic data. Our overall goal, said Mark Daly of the Broad Institute, Cambridge, Massachusetts, is to provide some biological insight that can be acted on therapeutically. So how do we move beyond the simple “stamp collecting” phase of cataloguing risk variants? Daly asked.
Figuring out how to make sense of all the variants identified to date will require “tremendous advances in high-throughput genomics and neurobiology that are not yet available,” he said. Although there is hope on the horizon, these advances are still in their early stages. Once we do get a handle on specific variants, Daly added, we will need to use model systems such as human cells and mice to produce relevant genetic changes that recapitulate, and ultimately reverse, the specific molecular changes that are caused by human disease variants. “Only then can we start thinking about moving beyond that,” he added.
Daly then shifted gears, focusing on how geneticists can use genome sequencing to inform biology. Sequencing is a great opportunity to look much further into the genetic underpinnings of diseases, he said. Rare and strong-acting variants can tell us more about what we should not target rather than what we should, he asserted. For example, several of the identified schizophrenia risk variants are located within genes in which severe loss-of-function mutations cause intellectual disability and other syndromic medical conditions. “These are probably not the genes we are looking to target with inhibitors,” Daly said.
Genetics can also be helpful in stratifying populations, said Daly, pointing to autism as an example. In cases of autism where general cognition is largely intact, there is no evidence of an excess of spontaneously arising mutations or copy number variants (CNVs); however, in cases with lower cognitive function, there is an enrichment of mutations affecting proteins that interact with fragile X mental retardation protein (FMRP), he said. Mutations in the FMRP gene lead to a form of intellectual disability called fragile X syndrome. These types of early genetic insights can be valuable, Daly suggested, because they hint that therapies designed around hypotheses of fragile X should be examined in specific autism populations.
The final speaker of the session was Tom Insel, director of the National Institute of Mental Health, Bethesda, Maryland, who said he was delighted to be able to attend the meeting, thanks to the end of the government shutdown just four days earlier. Compared to the state of the field back in 2005 (the last time the WCPG was held in Boston), we have made extraordinary progress, Insel said. “I think we have passed an inflection point,” he added.
Insel then described several points that he felt were the key take-home messages of the meeting that need to be conveyed to colleagues unable to attend. First, he said, genomics does not equal genes. The ENCODE project has demonstrated that there is a lot of information in the genome that is not related to protein coding sequences; in fact, some 20,000 non-coding RNA sequences have been identified. “Clearly, we have to rethink how we even conceptualize the genome,” he said. A second point is that genomics is not just the study of heritability. We now know that spontaneous mutations are very common, he stated, pointing to the average of 40 de novo mutations present in each individual that the 1000 Genomes Project has reported. Third, the genome is a “dynamic organ,” he said, and we are constantly generating variation. He cautioned that DNA from blood may not necessarily reflect the DNA in other tissues.
Insel posed two key questions: 1) When do you finally say, “We know enough about this variant,” such that it's worth spending huge amounts of money to gain a deeper understanding of what the variant does and how it’s related to disease? 2) Then, how do you go about getting that biological information? It’s important to look at how other diseases such as Alzheimer's and cancer are tackling this problem, he said, asserting that the problems facing psychiatric genetics are not different from those faced by other diseases.
In the works
Insel then turned to some of the ongoing projects that are attempting to answer these questions. NIH's Genotype-Tissue Expression (GTEx) Project is looking at how common variation affects gene expression in a variety of tissues, which will help to determine the variants that really matter in terms of function. In addition, high-throughput processes, such as TALENs and CRISPR, that manipulate genes of interest permit the examination of the effect of mutations on human biology. Insel also highlighted the importance of model animals in trying to reveal the fundamental biology of genetic variants.
In addition to the approach of going from phenotype to genotype (e.g., examining the genetic variants in people with schizophrenia), Insel said it is also important to move in the opposite direction: from genotype to phenotype (examining rare and common variants in the general population). An example of this approach is the UK Biobank project, which has already genotyped 500,000 individuals.
Evidence is mounting that mental illnesses are neurodevelopmental disorders, said Insel, highlighting the Allen Institute for Brain Science’s freely available BrainSpan as “a great resource for the community” that contains gene expression across fetal and postnatal development. The project has found that the pattern of gene expression in the fetal brain is profoundly different from that in the postnatal brain, said Insel. We think about psychiatric illnesses as circuit disorders, he said, and given that neural circuits are created during the second trimester, it is important to understand that circuits are developing in a molecular environment that is very different from that of adulthood.
Insel concluded his talk with a note of congratulations for the degree of data sharing in the field, but also an exhortation to continue data standardization, integration, and crowdsourcing in order to aggregate enough data. Bringing in other communities such as imaging and epidemiology is also critical to moving forward. If there is one thing that all four morning speakers would agree on, he said, it’s that these are still the early days and that there’s a long way to go.
The audience weighs in
Microphone lines were long as the session was opened up to questions and comments from the audience. In response to a complaint that the NIH isn’t providing enough funding for genomewide association studies, Insel said that a lot of money is being spent in that area, but acknowledged that “we just can’t buy all of the great science we want to” and added that private partners have stepped in to fill some of the gaps.
A suggestion was made to go back to the patients who have been genotyped to analyze phenotypic information. However, the fact that not all previous subjects have consented to be contacted again and there is a lack of standardization across groups in how phenotypic information is collected complicates matters. Still, Insel and Sullivan agreed that phenotypic information is important. “The PGC is banging one drum as loudly and quickly as possible,” said Sullivan, but added that there is a lot of value in other approaches. The Cross Disorders Group of the PGC is currently gathering available phenotypic information for its subjects, added session chair Jordan Smoller.
Much of the discussion focused on data sharing and team science. With the move toward consortia and the aggregation of data, single investigators become a small cog in the machine. What does this mean for trainees who will need to apply for grants as their careers move forward? “In some ways, [this new model] democratizes science,” said Insel, because it provides everyone with equal access to the data. Creativity becomes the rate-limiting step, he opined, so data sharing actually opens the field rather than closes it. As the way science is done changes, we also need to change how we recognize those who are involved in team science, Insel said, so that their important contributions are recognized.—Allison A. Curley.