NEW YORK – Investigators involved in the MedSeq Project have presented findings from an analysis of 100 participant genomes, taking into account the feasibility of analyzing and re-analyzing variants in thousands of genes.
Researchers from Partners Healthcare Personalized Medicine, Brigham and Women's Hospital, Harvard Medical School, the Broad Institute, and Massachusetts General Hospital assessed sequence data for 50 healthy individuals and 50 individuals with cardiomyopathy, focusing on more than 4,500 monogenic disease genes and variants that might serve as pharmacogenomic, blood antigen serotype, and/or polygenic risk markers. In the process, they initially identified apparent disease culprits in roughly half of individuals with cardiomyopathy.
As they reported online today in the American Journal of Human Genetics, the researchers also uncovered monogenic disease risk or carrier status in all but two of the 100 participants. Likewise, 95 percent of participants carried pharmacogenomics-related variants implicated in non-standard responses to one or more of the five drugs considered, and almost one-third of the individuals had at least one rare blood antigen genotype. More than a dozen variants were reclassified at the reanalysis stage of the study, which unearthed another 18 reportable variants.
"These findings highlight the quantity of medically relevant findings from a broad analysis of genomic sequencing data as well as the need for periodic reinterpretation and reanalysis of data for both diagnostic indications and secondary findings," corresponding and co-senior author Heidi Rehm, a molecular medicine, pathology, and medical and population genetics researcher affiliated with Partners Healthcare, BWH, Harvard, the Broad, and MGH, said in a statement.
In a paper published in BMC Medical Genetics in 2014, Rehm and colleagues shared medically relevant findings from the first 20 genomes analyzed for the MedSeq Project, and described the reporting procedures developed for those participants.
For the latest study, the team used standardized analyses to systematically search for clinically relevant variants in 100 genomes, considering variant- and gene-level results for more than 5,000 genes with potential ties to monogenic disease, carrier status, PGx traits, rare blood antigens, or complex-trait risk. The group subsequently explore the steps involved with reanalyzing and reporting new results anywhere from six months to almost two years after the initial analysis.
Starting with 5.3 million variants per genome, on average, the researchers narrowed in on just over 1,000 variants in each genome after filtering the data through their analytical pipeline.
They initially identified cardiomyopathy-related variants in 24 of the 50 participants with that condition, for example, but updated two of those cases as new information emerged on cardiomyopathy risky variants in the ALPK3 gene.
When it came to monogenic diseases, meanwhile, the team tracked down apparent monogenic disease contributors in 21 percent of participants after excluding cardiomyopathy-related variants. But 94 percent of the participants appeared to be monogenic disease carriers.
Reanalysis significantly bumped up the number of informative findings, uncovering new risk variants and/or reclassifying variants in 22 individuals, the authors reported, noting that "an ER-integrated interface enable automated real-time delivery of updates to physicians on previously reported variants."
Genome reanalysis has become a hot topic as more and more variant classifications come into focus. In a correspondence article in the New England Journal of Medicine last week, for example, researchers from Baylor College of Medicine and Baylor Genetics described new molecular diagnoses made by reanalyzing clinical exome sequence data generated between 2011 and 2013 using manual or semi-automated approaches.