Detection of CNVs from WES using ExomeDepth

CNV analysis was carried out using WES data of patients from our cohort who did not have a significant finding (i.e. SNV/indel) that could explain their phenotype.

CNV detection was performed using ExomeDepth (v. 1.1.15) [16]. For this analysis an in-house R script was implemented using R version 3.6 on the Cytera cluster of the Cyprus Institute (http://web.cytera.cyi.ac.cy/).

The analysis was performed with standard parameters using bam files created previously for SNV and indel variant calling procedure as described before. For each reference set, 5–10 samples were used depending on the available unrelated samples in the batch. For potential CNV calls on X chromosome only samples from the same sex were correlated.

A typical experiment produced on average 238 CNV calls. Results were filtered using Bayes factor (BF) >50 [17] and each call was further annotated using the online VEP tool (Ensembl release 81) and BioMart (http://grch37.ensembl.org/biomart/martview/).