At AMP 2014, Variant Interpretation in the Spotlight

AMP-logo

Whew! After last month’s AMP 2014 conference in National Harbor, Md., the QIAGEN Bioinformatics team could use a vacation. Between the steady interest in demos of our new Ingenuity Clinical Decision Support platform, the jam-packed exhibit hall booth, and attending world-class talks, the meeting was a veritable marathon.

We couldn’t be happier about how it all turned out, and it was a real honor to be able to show off our newest application to this savvy group of clinical lab professionals. The Ingenuity Clinical Decision Support platform was designed to accelerate variant interpretation and reporting of clinically relevant variants, so the AMP crowd was exactly the right audience for it.

At times it really felt like variant interpretation was the key theme of the entire conference. It popped up in all sorts of presentations, from those early-bird concurrent sessions at 7 am (ouch) to the plenary talks in front of nearly 2,000 attendees. During the opening keynote from NIH Director Francis Collins, he said that a major obstacle in getting to genomic medicine is our lack of knowledge about which genetic variants are associated with what outcome or disease. Delving into these variants, learning more about them, and accurately linking them to phenotypes are critical needs to achieve the community’s goal of precision medicine, Collins said, highlighting two public databases attempting to help with this: ClinVar and ClinGen.

In an award presentation, Uta Francke, professor emerita at Stanford University, spoke about her long-term research interest in mosaicism. She reported that filtering down a long list of variants to find causative ones is a crucial part of understanding unusual phenotypes — but cautioned against over-filtering, since variants that seem benign may turn out to be important in understanding certain disorders. She also predicted that whole genome sequencing will become widely available (both through the healthcare system and outside of it), and said she anticipates the development of an automated, machine-learning approach to deal with variant interpretation at the massive scale that will be required.

Eric Green, director of the National Human Genome Research Institute, talked about how genomics has become a big data field for the first time. Now, he said, the real bottleneck is data analysis; the community needs a robust, reliable way to quickly get a list of important variants and their functions from all of this sequence data. “We need systems to get us from variants to what they mean,” he said.

It was gratifying to see that this part of the genome interpretation process is important enough to have gotten the attention of the Genome in a Bottle consortium, an industry-wide group trying to nail down best practices for DNA sequencing. In a presentation from NIST’s Marc Salit, he said that the consortium aims to address everything from sample prep through variant calling and confidence estimates, eventually providing reference materials to ensure that everybody is getting comparable results with a high-quality benchmarking system. They’re also working with the Global Alliance (as are we) to support confidence calls in variant scoring for medical use.

AMP was a terrific experience for the QIAGEN Bioinformatics team, and we’re glad that our newest application could make a significant difference for clinical geneticists and other medical professionals as they work to achieve the goals outlined in presentations from some of the field’s greatest leaders.