Clinical exome studies have inconsistent coverage

BACKGROUND: Exome sequencing has become a commonly used clinical diagnostic test. Multiple studies have examined the diagnostic utility and individual laboratory performance of exome testing; however, no previous study has surveyed and compared the data quality from multiple clinical laboratories. METHODS: We examined sequencing data from 36 clinical exome tests from 3 clinical laboratories. Exome data were compared in terms of overall characteristics and coverage of specific genes and nucleotide positions. The sets of genes examined included genes in Consensus Coding Sequence (CCDS) (n = 17 723), a subset of genes clinically relevant to epilepsy (n = 108), and genes that are recommended for reporting of secondary findings (n = 57; excludes X-linked genes). RESULTS: The average exome nucleotide coverage (≥20×) of each laboratory varied at 96.49% (CV = 3%), 96.54% (CV = 1%), and 91.68% (CV = 4%), for laboratories A, B, and C, respectively. For CCDS genes, the average number of completely covered genes varied at 12 184 (CV = 29%), 11 687 (CV = 13%), and 5989 (CV = 37%), for laboratories A, B, and C, respectively. With smaller subsets of genes related to epilepsy and secondary findings, the CV revealed low consistency, with a maximum CV seen in laboratory C for both epilepsy genes (CV = 60%) and secondary findings genes (CV = 71%). CONCLUSIONS: Poor consistency in complete gene coverage was seen in the clinical exome laboratories surveyed. The degree of consistency varied widely between the laboratories.