RESUMEN
Glutaric acidemia I (GAI) is an autosomal recessive metabolic disease caused by a deficiency of glutaryl-CoA dehydrogenase enzyme (GCDH). Patients with GAI are characterized by macrocephaly, acute encephalitis-like crises, dystonia and frontotemporal atrophy. In this study, we investigated 18 Egyptian patients that were diagnosed with GAI based on their clinical, neuroradiological, and biochemical profiles. Of the 18 patients, 16 had developmental delay and/or regression, dystonia was prominent in 75% of the cases, and three patients died. Molecular genetics analysis identified 14 different mutations in the GCDH gene in the 18 patients, of the 14 mutations, nine were missense, three were in the 3'-Untranslated Region (3'-UTR), one was nonsense, and one was a silent mutation. Four novel mutations were identified (c.148 T > A; p.Trp50Arg, c.158C > A; p.Pro53Gln, c.1284C > G; p.Ile428Met, and c.1189G > T; p.Glu397*) that were all absent in 300 normal chromosomes. The 3'-UTR mutation (c.*165A > G; rs8012), was the most frequent mutation observed (0.5; 18/36), followed by the most common mutation among Caucasian patients (p.Arg402Trp; rs121434369) with allele frequency of 0.36 (13/36), and the 3'-UTR mutation (c.*288G > T; rs9384, 0.22; 8/16). The p.Arg257Gln mutation was found with allele frequency of ~0.17 (6/36). The marked homozygosity observed in our patients is probably due to the high level of consanguinity that is observed in 100% of the cases. We used nine in silico prediction tools to predict the pathogenicity (SIFT, PhD-SNP, SNAP, Meta-SNP, PolyPhen2, and Align GVGD) and protein stability (I-Mutant2.0, Mupro, and istable) of the nine missense mutants. The mutant p.Arg402Trp was predicted to be most deleterious by all the six pathogenicity prediction tools and destabilizing by all the three-stability prediction tools, and highly conserved by the ConSurf server. Using the clinical, biochemical, family history of the 18 patients, and the in silico analysis of the missense mutations, our study showed a mix of conclusive and inconclusive genotype-phenotype correlations among our patient's cohort and suggests the usefulness of using various sophisticated computational analysis to be utilized for future variant classifications in the genetic clinics.