Suzanne de Bruijn

236 Chapter 5 MIPs were designed to cover exons and exon-intron boundaries of a panel of 120 HL genes ( Table S3 ). For each targeted region an average coverage of >500 reads was obtained. Aminimal coverage of 20x was reached for 91.78% of the MIPs. CNV detection for SLC26A4 was performed using a read coverage analysis as previously described. 34 . Additionally, coding and splice site regions of FOXI1 and the regions harboring reported pathogenic variants in EPHA2 were sequenced using Sanger sequencing as previously described 35 , since these genes are not included in the MIP panel. Primer sequences and PCR conditions are available upon request. Variant prioritization was based on an AF of ≤0.5% (gnomAD V2.1.1 36 and our in-house exome database (~15,000 alleles)), unless specified otherwise. Variant visualization was performed using the IGV software V.2.4 (Broad Institute). 37 Interpretation of missense variants was performed using the in silico tools CADD-PHRED (≥15) 38 , SIFT (≤0.05) 39 , PolyPhen-2 (≥0.450) 40 and MutationTaster (deleterious) 41 to predict potentially deleterious effects. Variants were prioritized if a deleterious effect was predicted by at least two of these tools. Candidate variants were validated by Sanger sequencing and segregation analysis was performed when DNA of family members was available. Primer sequences and PCR conditions are available upon request. Potential effects on splicing of missense, synonymous and intronic variants were assessed using the deep- learning splice prediction algorithm SpliceAI (≥0.1). 42 The maximum distance between the variant and potential gained or lost splice sites was set to 1000 bp. Predicted splice altering defects were evaluated using an in vitro splice assay in HEK293T cells as previously described. 43 Detection of the CEVA haplotype Initial identification of the CEVA haplotype 22 was performed with SNP-genotyping by Sanger sequencing in index cases for whom parental DNA was available for segregation analysis. Subsequently, the corresponding VNTR marker haplotype was determined in CEVA-positive families. For additional cases, VNTR marker analysis was performed to enable a fast and cost-effective detection of the CEVA haplotype. For the VNTR marker analysis, DNA segments were amplified by employing touchdown PCR, and subsequent analysis was carried out on an ABI Prism 3730 Genetic Analyzer (Applied Biosystems). Genomic positions of the markers were determined using the UCSC genome browser (GRCh37/hg19). 44 Alleles were assigned with the GeneMarker software (V.2.6.7, SoftGenetics) according to the manufacturer’s instructions. When an individual was suspected of carrying the CEVA haplotype based on VNTR-marker alleles,