Materials and methods

The current study tested 24 unrelated patients aged <16 years who presented to the Paediatric Cardiomyopathy Clinic of Cairo University Children’s Hospital with primary idiopathic hypertrophic cardiomyopathy. Exclusion criteria included patients older than 16 years, infants of diabetic mothers, patients with secondary causes such as systemic hypertension, aortic stenosis (valvular and sub-valvular), infiltrative cardiomyopathies, storage disorders, and other metabolic causes or intake of drugs causing cardiac hypertrophy. Syndromic patients and patients with extracardiac manifestations were also excluded. Included cases were subjected to the following clinical and genetic evaluations:

1. 1. Clinical workup: A) A complete medical history including antenatal, natal histories, any maternal chronic illness, exposure to drug intake, and history of cardiac symptoms especially those favouring the diagnosis of cardiomyopathy, such as dyspnea, palpitation, recurrent chest infections, and syncopal attacks, B) A full family history along with a three-generation pedigree that includes information on consanguinity, similar conditions in the same family or other genetic diseases. Familial hypertrophic cardiomyopathy was considered if at least one of the family members of the proband was diagnosed with the disease, ^{Reference Bottillo, D’Angelantonio and Caputo33} C) Anthropometric measurements (e.g., weight, height, and head circumference), D) Full clinical examination, E) Measurement of the blood pressure to exclude systemic hypertension, F) Plain chest X-ray, and G) Electrocardiogram and echocardiography (M-mode and Doppler). Electrocardiogram and echocardiography were done for the cases and parents as well. Echocardiographic parameters included left ventricular end-systolic dimension, left ventricular end-diastolic dimension, maximum left ventricular wall thickness, interventricular septal thickness, left ventricular fractional shortening, left ventricular ejection fraction, and left ventricular outflow tract pressure gradient. Patterns of hypertrophy were also reported. ^{Reference Elliott and McKenna34} The left ventricular diastolic wall thickness > 2 SD from the predicted mean was the base of diagnosis of hypertrophic cardiomyopathy (z-score > 2, and z-score is defined as the number of SD from the body surface area adjusted mean in the normal population). ^{Reference Simpson and Chubb35} Hypertrophic obstructive cardiomyopathy was considered if the left ventricular outflow tract pressure gradient is more than 30 mmHg, ^{Reference Maron, Olivotto and Betocchi36} H) Heart failure was assessed according to ROSS/NYHA functional classification, ^{Reference Yancy, Jessup and Bozkurt37} I) Other investigations were performed as appropriate, for example, for the exclusion of inborn errors of metabolism presenting with cardiomyopathy, laboratory assessment included routine lab investigations (complete blood picture, liver and kidney function tests, electrolytes, creatine kinase, and blood glucose), and specialised lab investigations (extended metabolic screen for the exclusion of amino acid, fatty acid and organic acid disorders, and enzymatic assays for lysosomal and storage diseases).

2. 2. Genetic Testing. Targeted enrichment next-generation sequencing was performed on Illumina MiSeq system using TruSight Cardio panel (Illumina, San Diego, CA, USA). ^{Reference Pua, Bhalshankar and Miao38} This panel contains 174 genes with known associations to 17 inherited cardiac conditions.

DNA Extraction and quantification. Genomic DNA was extracted using a DNA extraction kit (NucleoSpin Blood QuickPure, Macherey Nagel GmbH & Co., Germany). The DNA was quantified using Qubit® dsDNA HS assay kit (Invitrogen, Grand Island, NY, USA) and the Qubit 2.0 Fluorimeter (Thermo Fisher Scientific, CA, USA).

Library preparation. Target regions were captured by in-solution hybridisation according to the manufacturer protocol. The targeted regions were sequenced using the MiSeq platform, generating two million of 150 bp paired-end reads for each sample (Q30 ≥ 90%).

Bioinformatic data analysis and variant classification. The MiSeq Reporter software installed on-instrument was used for demultiplexing the data and raw FASTQ file generation. Bam files were made using the Burrows–Wheeler Aligner ^{Reference Li and Durbin39} which aligned the reads against the human reference genome GRCh37/hg19. Variants were called using the Genome Analysis Toolkit ^{Reference Van der Auwera, Carneiro and Hartl40} and variant call format files were created. A run is called accepted if at least 95% of the target regions were covered with an average depth of coverage 100X. Integrated Genome Viewer 2.4 was used to visually inspect sequence reads and variant positions. ^{Reference Thorvaldsdóttir, Robinson and Mesirov41} A variant was accepted if it had a quality score > 100 and covered by at least 50 reads. The variant call format files were then annotated using in-house developed scripts based on the ANNOVAR knowledge database. ^{Reference Wang, Li and Hakonarson42} The annotation included information about the physical position of the variant in the genome and the protein using SnpEff (v4.3 T) ^{Reference Cingolani, Platts and Wang43} (http://snpeff.sourceforge.net/) and the minor allele frequency of variants in public databases such as Genome Aggregation Database (http://gnomad.broadinstitute.org/), 1000 genome, and dbSNP (V138) (www.ncbi.nlm.nih.gov/SNP/). The effect of the variants on the protein structure was predicted using bioinformatics’ algorithms such as Sorting Tolerant From Intolerant (http://sift.jcvi.org/), PolyPhen-2 (http://genetics.bwh.harvard.edu/pph2/), and Combined Annotation Dependent Depletion (CADD). ^{Reference Rentzsch, Witten, Cooper, Shendure and Kircher44} Clinical annotation was performed using Online Mendelian Inheritance in Man (https://omim.org), Human Gene Mutation Database (HGMD Professional 2019.4; http://www.biobase-international.com/), and ClinVar ^{Reference Landrum, Lee and Riley45} (http://www.ncbi.nlm.nih.gov/clinvar/).

The variants were filtered to exclude non-relevant variants (e.g., synonymous and intronic variants). Variants with minor allele frequency >0.01% in population databases were considered frequent and excluded. Exonic variants predicted to be tolerant/benign by Sift/Polyphen/CADD were also excluded. The remaining variants were evaluated by investigating their records in Online Mendelian Inheritance in Man, Human Gene Mutation Database, and ClinVar. The panel included 15 recessive genes related to cardiomyopathy (ALMS1, COX15, DMD, FKRP, FKTN, FXN, GAA, GATAD1, HADHA, LAMA2, SCO2, SDHA, SGCB, SGCD, and SGCG). Online Mendelian Inheritance in Man was used to check the inheritance model of the gene, and a variant not matching the model was excluded. Specifically, a heterozygous variant in a gene whose inheritance model was autosomal-recessive was considered a carrier and not considered the disease cause. The pathogenicity of the variant was evaluated using the Human Gene Mutation Database and ClinVar, any variant classified as benign was also excluded. The remaining variants were examined by extensive literature search to evaluate their contribution to the disease. Finally, the variants were classified according to the American College of Medical Genetics and Genomics guidelines ^{Reference Richards, Aziz and Bale46} as a pathogenic, likely pathogenic, variant of uncertain significance, benign or likely benign.

The identified variants were confirmed by Sanger sequencing using the ABI 3730XL DNA analyzer (Applied Biosystems, Foster, CA, USA).