This is an excerpt of a story that appeared in Science. Read the full article here.
Personal DNA sequencing once promised to up the ante for individualized medicine. Perhaps no one believed that more than human genomics pioneer J. Craig Venter, who in 2014 co-founded a company called Human Longevity to predict and prevent disease by sequencing a million human genomes. But Venter is no longer content with your DNA. His latest venture—a subsidiary called Health Nucleus based in San Diego, California—says it can detect undiagnosed health problems by combining DNA analyses with a $25,000 workup including a whole-body MRI scan, metabolomics screening, 2 weeks of constant heart monitoring, pedigree analysis, microbiome sequencing, and a glut of standard laboratory tests.
Precision Medicine Initiative—is the way of the future. But many other clinicians and researchers are leery or even downright outraged by the program’s potential for over diagnosis and what they see as lack of evidence for its benefits.Enthusiasts of “precision medicine” say this kind of screening—similar to the U.S. National Institutes of Health’s (NIH’s)
Late last week, Venter and co-workers quietly published a paper on the preprint server bioRxiv—which does not use peer review—that presents data from the new project. According to the study, screening detected “age-related chronic diseases requiring prompt (<30 days) medical attention” in 8% of the 209 participants, and MRIs found early-stage cancer in 2%. However, Health Nucleus did not confirm that the data were from its $25,000 medical exam, although descriptions of the diagnostics were nearly identical.
“It’s a classic Craig Venter study that pushes the envelope of what is considered reasonable,” says Olivier Elemento, associate director at the Institute for Computational Biomedicine at Weill Cornell Medicine in New York City.
Despite Venter’s personalized genomics evangelism, the study’s results pointedly indicate that “the genome alone doesn’t tell you the whole story,” says Elemento of Cornell. Only 25% of patients had probable links between gene variants and disease phenotypes. “But when you can combine genes with an additional readout that tells you the gene is doing something, your ability to predict disease increases dramatically.”