Genome-scale sequencing in clinical care: establishing molecular diagnoses and measuring value.

Abstract

The arrival of genomic medicinewas presaged by the publication of the draft human genome sequence in 2001.1,2 Now, the application of genome-scale sequencing is being rapidly adopted in academic centers in cases for which numerous individual genetic tests would be required for evaluation, the condition is genetically heterogeneous, or genetic testing is simply not available to evaluate the suspected diagnosis. Developmental delay in children is an example of an indication for such testing. In this issue of JAMA, 2 groups provide compelling evidence for the ability of exome sequencing to establish a molecular diagnosis.3,4 Exome sequencing evaluates the approximately 1% of the genome that encodes proteins, which is where the majority of variants that cause monogenic disorders are located. These results are different than a previous study published in JAMA that examined the application of whole-genome sequencing among 12 healthy adults.5 Whole-genome sequencing detects variations in both protein-coding regions as well as the spaces between genes in the genome. Interpretation of the functional consequences of sequence variations between genes is very much a scientific frontier and fraught with challenges. Another substantial difference is that of prior probability: the likelihood that a healthy individual in the general population is affected by a rare genetic disorder is extremely small, whereas every patient reported in the current studies had symptoms and clinical findings consistent with a genetic etiology. The study by Yang and colleagues3 reported findings from 2000 consecutive patients who had clinical wholeexome sequencing performed at a single academic clinical genetics laboratory over 2 years (from June 2012 to November 2013). The study population was mostly children (n = 1756; mean age, 6 years) and exhibited diverse clinical manifestations, most commonly involving nervous system dysfunction such as developmental delay. The authors report that a molecular diagnosis was established for 504 patients (25.2%) and that medically actionable findings were present in 92 patients (4.6%). The authors also provided information on a summary analysis of unselected, unrelated cases completed and reported from the close of the current 2000 case cohort (November 2013) through August 30, 2014, bringing the total number of cases included in this report to 3386 patients. The overall diagnostic rate for the total cases remained unchanged at 25% (830 molecular diagnoses out of 3386 total cases). In the study by Lee and colleagues,4 clinical exome sequencing (CES) was performed on 814 consecutive patients (64% [n = 520] were children aged ≤18 years, of whom half [n = 254] were aged <5 years) at a single academic clinical genomics center from January 2012 and August 2014, with testing involving trio-CES (typically simultaneous sequencing of both parents and their affected child) or proband-CES (including sequencing of only the affected individual) when parental samples were not available. The authors report that, overall, a molecular diagnosis was established for 213 patients (26%), with rates of molecular diagnosis of 31% (127 of 410 cases) for trio-CES and of 22% (74 of 338 cases) for proband-CES. Together, the studies represent more than 2800 clinical cases spanning a broad spectrum of patient ages and conditions. The overall diagnostic yields reported by these groups are remarkablysimilardespitedifferences in technical andanalytical approaches. Lee et al report a significantly higher diagnostic yield when sequencing trios, largely due to an enhancedability todetect denovoand compoundheterozygous variants. In both studies, the authors noted examples of positive findings related to genes that had been recently implicated in genetic disorders, highlighting the flexibility of the exome platform to incorporate new knowledge in real time. Thepatients reported inthesearticleshadsymptomsthat touch on nearly every medical specialty, although the data suggest ahigheryield incases involvinganeurologicalphenotype,particularly in children. However, it is difficult to discern the overall performance (clinical sensitivity and clinical specificity) of exome sequencing from these data. Some of the positive findings, hopefully very few in number, will be false-positive results due to misinterpretation of variant pathogenicity or misattribution of gene-disease association in the medical literature. Many of the negative findings are likely false-negative results, with genetic etiologies undetectable by exome technology, unknown to current medical science, or both. In some cases the presenting condition may not be genetic in origin, in which case the negative findings would be true negative results. Further details about the performance of exome sequencing will be needed as its use as a routine clinical test is contemplated. What is clear from these studies is that powerful new genetic sequencing technologies canelucidate a substantial percentage of cases, many of which were previously impenetrable with existing diagnostic modalities. How will this information affect care?Whichpatientswill benefit themost? Related articles pages 1870 and 1880 Opinion