Exome sequencing in suspected monogenic dyslipidemias

Abstract

Background: Exome sequencing is a promising tool for gene mapping in Mendelian disorders. We used this technique in an attempt to identify novel genes underlying monogenic dyslipidemias.

Methods and results: We performed exome sequencing on 213 selected family members from 41 kindreds with suspected Mendelian inheritance of extreme levels of low-density lipoprotein cholesterol (after candidate gene sequencing excluded known genetic causes for high low-density lipoprotein cholesterol families) or high-density lipoprotein cholesterol. We used standard analytic approaches to identify candidate variants and also assigned a polygenic score to each individual to account for their burden of common genetic variants known to influence lipid levels. In 9 families, we identified likely pathogenic variants in known lipid genes (ABCA1, APOB, APOE, LDLR, LIPA, and PCSK9); however, we were unable to identify obvious genetic etiologies in the remaining 32 families, despite follow-up analyses. We identified 3 factors that limited novel gene discovery: (1) imperfect sequencing coverage across the exome hid potentially causal variants; (2) large numbers of shared rare alleles within families obfuscated causal variant identification; and (3) individuals from 15% of families carried a significant burden of common lipid-related alleles, suggesting complex inheritance can masquerade as monogenic disease.

Conclusions: We identified the genetic basis of disease in 9 of 41 families; however, none of these represented novel gene discoveries. Our results highlight the promise and limitations of exome sequencing as a discovery technique in suspected monogenic dyslipidemias. Considering the confounders identified may inform the design of future exome sequencing studies.

Keywords: DNA sequencing; exome; genetics; human; lipids.

Figure 1. Selected metrics from exome sequencing analysis

(a) Percent targeted bases across the exome definition supported by ≤ 20 sequencing reads (Red) or ≤ 10 sequencing reads (Blue) or across genes previously identified to cause monogenic dyslipidemia supported by ≤ 10 sequencing reads (Green). (b) Number of candidate variants after analysis for families with a suspected pathogenic variant in a gene known to cause monogenic dyslipidemia (orange) compared with families without known cause (brown). P refers to the p-value from the Kolmogorov–Smirnov test in testing for differences between the distributions.

Figure 2. Pedigree of Family A7, demonstrating the utility of refining phenotypes based on burden of common alleles

(A) Initial pedigree defining affected individuals (shaded) by LDL-C level adjusted for age and gender. (B) Updated pedigree based on residualized phenotype score (see text) where individuals III-1 and III-2 are classified as unaffected. LDLR p.E228K carrier status is indicated with + (heterozygous) or - (wild type). The superscript (a) indicates the LDL-C level was obtained while the individual was on lipid-lowering medication therapy.

Figure 3. Discovery rates from exome sequencing

The distribution of final discovery status for the 41 families with suspected monogenic dyslipidemias that underwent exome sequencing is shown with approximate percentages.