Defining quality standards for clinical whole exome sequencing: a national collaborative study of the Dutch Society for Clinical Genetic Laboratory Diagnostics (VKGL)

, , , , , , , , , , , , , ,

Research output: Contribution to conferenceAbstractOther research output

Abstract

Clinical whole exome sequencing (WES) has proven to be very effective for diagnosing heterogeneous genetic diseases, and has therefore been adopted in standard Dutch genetic diagnostic services. Quality standards have been described for gene panel based next generation sequencing, but for clinical WES these have not been determined yet. In this study, we present a nationwide quality assessment scheme within all eight university medical centers in the Netherlands in order to improve and harmonize quality and to formulate a quality standard for clinical WES.

The exome of Genome in a Bottle (GIAB) sample NA12878 was sequenced and processed by the Dutch genetic diagnostic centers according to their own standard diagnostic protocols. Anonymized VCF and BAM files were collected and used for comparisons. VCF files were used to calculate variant detection sensitivity and precision using the GIAB high confidence call set (v2.19). BAM files were used to calculate informative exome coverage statistics based upon a standardized target defined by all coding exons of UCSC and Ensembl +/- 20bp intron flanks.

Results showed that single nucleotide variant (SNV) detection sensitivity varied between 96.3% and 99.1% (average 98.1%) and that INDEL detection sensitivity varied between 86.1% and 96.1% (average 90.4%). Precision of SNV detection was between 98.0% and 99.8% (average 98.9%), and precision for INDEL detection was between 74.8% and 97.0% (average 91.0%). The mean coverage of the exome varied between 55X and 152X. Completeness of the Exome (defined as the percentage of bases with 15x informative coverage) varied between 90.0% and 96.4% (average 93.5%). Downsampling analysis indicated that completeness is a good predictor for quality and can be used to determine the required sequencing output for a desired quality threshold.

The nationwide quality assessment scheme shows an acceptable concordance between the Dutch genetic diagnostic centers, although harmonisation of lab protocols and bioinformatics may lead to improved concordance and quality. We propose the following quality standards for a clinical WES based upon the GIAB high confidence call set (v2.19): a minimum sensitivity of 97% for SNVs and 85% for INDELs, and a minimum precision of 98% for SNVs and 85% for INDELs. These quality assurance standards will be deposited as field standards at the Dutch accreditation council for use in ISO15189 laboratory accreditation.

Original languageEnglish
Publication statusPublished - 22 Sept 2017
EventNVHG Najaarssymposium - NH Eindhoven Conference Centre Koningshof, Veldhoven, Netherlands
Duration: 21 Sept 201722 Sept 2017
http://www.nvhg-nav.nl/page.aspx?page=congresinfo

Conference

ConferenceNVHG Najaarssymposium
Country/TerritoryNetherlands
CityVeldhoven
Period21/09/1722/09/17
Internet address

Fingerprint

Dive into the research topics of 'Defining quality standards for clinical whole exome sequencing: a national collaborative study of the Dutch Society for Clinical Genetic Laboratory Diagnostics (VKGL)'. Together they form a unique fingerprint.

Cite this