Estimating diagnostic noise in panel-based genomic analysis

Abstract

Background: Gene panels with a series of strict variant filtering rules are often used for clinical analysis of exomes and genomes. Panels vary in size, which affects the sensitivity and specificity of the test. We sought to investigate the background rate of candidate diagnostic variants in a population setting using gene panels developed to diagnose a range of heterogeneous monogenic diseases. Methods: We used the Genotype-2-Phenotype database with the Variant Effect Predictor plugin to identify rare non-synonymous variants in exome sequence data from 200,643 individuals in UK Biobank. We evaluated five clinically curated gene panels: developmental disorders (DD; 1708 genes), heritable eye disease (536 genes), skin disorders (293 genes), cancer syndromes (91 genes) and cardiac conditions (49 genes). We further tested the DD panel in 9,860 proband-parent trios from the Deciphering Developmental Disorders (DDD) study. Results: As expected, bigger gene panels resulted in more variants being prioritised, varying from an average of ~0.3 per person in the smallest panels, to ~3.5 variants per person using the largest panel. The number of individuals with prioritised variants varied linearly with coding sequence length for monoallelic disease genes (~300 individuals per 1000 base pairs) and quadratically for biallelic disease genes, with some notable outliers. Based on cancer registry data from UK Biobank, there was no detectable difference between cases and controls in the number of individuals with prioritised variants using the cancer panel, presumably due to the predominance of sporadic disease. However, we observed a marked increase in the number of prioritised variants in the DD panel in the DDD study (~5 variants per proband). Phasing of compound heterozygotes in biallelic genes resulted in a modest reduction in the number of prioritised variants. Conclusions: Although large gene panels may be the best strategy to maximize diagnostic yield in genetically heterogeneous diseases, they will frequently prioritise false positive candidate variants potentially requiring additional clinical follow-up. Most individuals will have at least one rare nonsynonymous variant in panels containing >500 monogenic disease genes. Extreme caution should therefore be applied when interpreting potentially pathogenic variants found in the absence of relevant phenotypes.

Genomics and Insurance in the United Kingdom: Increasing Complexity and Emerging Challenges

“…Indeed, large numbers of variants that have historically been considered to be pathogenic (associated with specific health outcomes) have in fact turned out to be common in individuals who do not show the associated phenotype (36), suggesting that either their original classification was wrong, or that their impact on health is more subtle or contextdependent than previously appreciated. Beaumont et al (37) illustrated the challenge of interpreting people's genomic data, showing that while large gene panels may maximize diagnostic yield, they are also likely to identify several variants that look hypothetically concerning though are probably benign: most people have at least one rare variant in the coding regions of the genome in panels containing over five hundred disease genes. Even for 'well-understood' pathogenic genetic variants, context matters: Jackson et al (38) found that people with cancer-predisposing genetic variants were at significantly less elevated risk of cancer in the absence of a family history.…”

PhenoScore: AI-based phenomics to quantify rare disease and genetic variation

“…Clinical WES data interpretation relies on filtering and prioritization for rare genetic variants in disease-gene panels, which are subsequently interpreted in the context of the patient's clinical presentation [5]. Whereas this strategy is essential to identify the disease-causing variant(s), it is estimated that, depending on the number of genes included in the panel, dozens of variants are prioritized as diagnostic noise [6] -and this number is expected to rise even more in the coming years with technological innovations such as genome sequencing finding their way into the diagnostic arena [7][8][9].…”