A survey of undetected, clinically relevant chromosome abnormalities when replacing postnatal karyotyping by Whole Genome Sequencing

Abstract

Whole genome sequencing (WGS) holds the potential to identify pathogenic gene mutations, copy number variation, uniparental disomy and structural rearrangements in a single genetic test. With its high diagnostic yield and decreasing costs, the question arises whether WGS can serve as a single test for all referrals to diagnostic genome laboratories ("one test fits all"). Here, we provide an estimate for the proportion of clinically relevant aberrations identified by light microscopy in postnatal referrals that would go undetected by WGS. To this end, we compiled the clinically relevant abnormal findings for each of the different referral categories in our laboratory during the period 2006-2015. We assumed that WGS would be performed on 300-500 bp DNA fragments with 150-bp paired sequence reads, and that the mean genome coverage is 30x, corresponding to current practice. For the detection of chromosomal mosaicism we set minimum thresholds of 10% for monosomy and 20% for trisomy. Based on the literature we assumed that balanced Robertsonian translocations and ∼9% of other, balanced chromosome rearrangements would not be detectable because of breakpoints in sequences of repetitive DNA. Based on our analysis of all 14,957 referrals, including 1455 abnormal cases, we show that at least 8.1% of these abnormalities would escape detection (corresponding to 0.79% of all referrals). The highest rate occurs in referrals of premature ovarian failure, as 73.3% of abnormalities would not be identified because of the frequent occurrence of low-level sex chromosome mosaicism. Among referrals of recurrent miscarriage, 25.6% of abnormalities would go undetected, mainly because of a high proportion of balanced Robertsonian translocations. In referrals of mental retardation (with or without multiple congenital anomalies) the abnormality would be missed in only 0.35% of referrals. These include cases without imbalances of unique DNA sequences but of clinical relevance, as for example, r(20) epilepsy syndrome. The expected shift to large-scale implementation of WGS ("one test fits most") as initial genetic test will be beneficial to patients and their families, since a cause for the clinical phenotype can be identified in more cases by a single genetic test at an early phase in the diagnostic process. However, a niche for genome analysis by light microscopy will remain. For example, in referrals of newborns with a suspicion of Down syndrome, karyotyping is not only a cost-effective method for providing a quick diagnosis, but also discriminates between trisomy 21 and a Robertsonian translocation involving chromosome 21. Thus, when replacing karyotyping by WGS, one must be aware of the rates and spectra of undetected abnormalities. In addition, it is equally important that requirements for cytogenetic follow-up studies are recognized.

Keywords: Chromosomal mosaicism; Diagnostic yield; Karyotyping; Whole genome sequencing.