It is great to be here in Marco Island for the AGBT meeting! The 16th annual meeting hit the ground running with a pre-meeting workshop hosted by the Genome Reference Consortium (GRC) followed by an opening session that was more clinically focused than many attendees are used to at this tech-heavy conference. From the dynamic Q&A sessions, it was clear that these were precisely the kind of talks that people have been looking for as this meeting evolves downstream along with genomic science.
The GRC workshop, entitled ‘Advancing the Human Reference Assembly’ included four speakers: Valerie Schneider (NCBI), Tina Graves-Lindsay (TGI), Karyn Meltz Steinberg (TGI) and Deanna Church (Personalis, Inc.) They stated that the current human genome reference assembly represents a mixture of over 70 individuals’ genomes in a single linear sequence. Thanks to population sequencing efforts, like the 1,000 Genome Project, we now know that there are regions of the human genome that are highly polymorphic, with multiple haplotypes that are segregating in the global population. Many of these regions (like MHC and KIR) are heavily associated with disease and immune-response.
Making accurate genotype calls in these regions, which include a large degree of structural variation, requires improved reference sequence representation of these population-specific haplotypes. Deanna Church commented on the significance of this by saying, “If you want to do genotype and phenotype association, you better get genotype correct.” To better represent these population-specific haplotype reference sequences, the GRC’s latest human genome build (GRCh38) includes alternative loci. With the inclusion of these alt loci, Valerie Schneider presented how we’re now entering the “Multiple Genome Era”, where the human reference genome will be expanded to include additional alt loci sequence to better represent the population diversity of these highly polymorphic loci.
To sequence and assemble these new reference sequences, Tina Graves-Lindsay and Karyn Meltz Steinberg then outlined a new strategy to build gold and platinum quality genome assemblies, using long-read data (PacBio® data) and other complimentary technologies. The GRC presenters also described how the GRC will be adding more alt loci diversity to the Human Genome Reference by using PacBio sequencing to do whole-genome de novo sequencing and assembly of additional individuals from under-represented populations. Check out the slides from workshop on slideshare.
The formal session began with a talk from David Goldstein, who recently moved from Duke University to Columbia University, about precision medicine in neurological disease. He focused on several large-scale studies of patients with epilepsy, including a project that sequenced more than 350 affected children and their parents to find de novo mutations. While many of the mutations detected are rare, Goldstein noted that they often are found in common biological pathways so it may be possible to stratify patients. In one example of such grouping, patients with a mutation in the KCNT1 gene were given quinidine, an FDA-approved drug that was never previously indicated for epilepsy. In three cases, two patients saw significant improvement in the severity or frequency of their seizures. Goldstein pointed out that this targeted approach would never have been found without genomics.
As he presented similar examples from other studies, Goldstein noted that the stakes are considerably higher for getting the genetics right when the goal is diagnostic sequencing. He also said that patients’ genomes need to be comprehensively interrogated; in epilepsy, at least, new variants are being found so often that a gene panel approach wouldn’t keep up. His talk was hopeful for the future of clinical sequencing and improved bioinformatics to explain findings: results are already impressive, he said, but “this is only going to get better.”
Rick Lifton from Yale University also spoke in the kickoff session, focusing on the need to determine the function of more of the human genome than is currently understood. “There’s an awful lot of room for new discovery,” he said, pointing out that of the 21,000 known protein-coding genes in the human genome, only 3,000 have been clearly linked to disease. Lifton discussed studies of various conditions, such as hypertension, where typical approaches for understanding Mendelian disease proved useful for more common diseases. Based on success in finding de novo mutations in several studies, Lifton called for routine sequencing in the clinic. He added that truly understanding the human genome will require elucidating noncoding regions, determining the consequence of every mutation, and identifying biological targets for therapeutics.
The rest of the agenda looks just as interesting. We’ll keep reporting from the sessions, so check back for more soon!