Abstract Introduction Barth Syndrome is an X-linked multisystem disease caused by the mutation in the TAZ gene (TAZ, G 4.5, OMIM 300394) that encodes for acyltransferase taffazin. Taffazin is a mitochondrial protein highly expressed in heart and skeletal muscle and has a central role in cardiolipin remodeling process. TAZ is the only known single gene defect in cardiolipin remodeling, with more than a hundred TAZ mutations identified. However, no systematic correlation has been established between TAZ genotypes and clinical or biochemical phenotypes in Barth syndrome. There is a large phenotypic heterogeneity among affected individuals sometimes even within the same family. This raises a possibility of genetic and metabolic modifiers involved in TAZ clinical and metabolic phenotypes. The aim of this study is to apply LC-MS/MS untargeted and targeted metabolic analyses to explore possible correlations between two TAZ genotypes and metabolome in a cellular model of Barth Syndrome. Methods Mutant TAZ 3(c.647G>T), TAZ 4(c.778-24_778-7delinsA) fibroblast and matching by donors’ age and gender control fibroblast lines (healthy#1, healthy #2) were cultured under identical conditions. Cells (n = 6 for each genotype) were harvested and extracted followed Bligh and Dyer protocol. Extracted metabolites were analyzed by Agilent 6545 Q-TOF coupled with 1290 Infinity II HPLC. Chromatographic separation was achieved by using a Water Premier BEH Z-HILIC Column (2.5 µM, 2.1 × 150 mm) and Waters X-select HSS T3 (2.5 µM, 2.1 × 100 mm) columns for polar and non-polar metabolites respectively. Raw data was normalized with internal standard and total cellular protein amounts. Statistical and pathway enrichment analyses were performed with a chemometric platform Mass Profiler Professional (Agilent) and Metaboanalyst webtool. Results Data was treated first as for two separate genotypes (healthy vs TAZ) followed and compared to the analysis with four separate genotypes (healthy#1, healthy #2 TAZ c.647G>T, TAZ c.778-24_778-7delinsA). Metabolites contributing to discrimination of the genotypes based on metabolic profiles were identified by using a univariate analysis (P < 0.05) and a supervised PLSDA (VIP>1 for small polar metabolites and VIP > 3 for non-polar molecules). The differentially expressed metabolites included amino acids, acylcarnitine, Krebs cycle intermediates, ATP, 3-methylglutaconic acid, fatty acids and a variety of phospholipids. Pathway enrichment analysis revealed that several metabolic pathways were involved. Conclusions Our analyses reveal that genetic variations in TAZ gene cause the differences in the abundance of certain metabolites and affected pathways. This leads us to the conclusion that from the point of view of metabolic pathophysiology, TAZ should be considered beyond a single-gene defect and a more personalized approach is needed in search for potential therapies.