RESUMO
We present a GC-MS metabolomics workflow for analyzing metabolites in urine samples infected with schistosomiasis. Schistosomiasis, a neglected tropical disease, affects 85% of the global population, with the majority residing in Sub-Saharan Africa. The workflow utilized in this study involved the utilization of the AMDIS freeware, Metab R for pre-processing, and multivariate statistical classification through partial least squares-discriminant analysis (PLS-DA). This classification aimed to categorize volatile metabolites found in urine samples from humans infected with schistosomiasis. All samples were collected from individuals in Botswana. A solid-phase microextraction-fused silica fiber was used to adsorb volatile metabolites from the urine samples and inserted into the GC-MS injection port for data acquisition. The acquired data were then subjected to AMDIS auto-deconvolution, Metab R pre-processing, and statistical evaluation for metabolite mining. A total of 12 metabolites, including 3-chloropropionic acid and heptadecyl ester with an AMDIS match factor of 96% at an approximated amount of 0.35% and cyclohexylamine with an AMDIS match factor of 100% and approximated amount of 0.39%, were identified. PLS-DA was used for the classification of the metabolites. The method showed good sensitivity and specificity as indicated by the receiver operating characteristic measured by the areas under the curves. Results indicated that metabolomics is a useful tool for mining metabolites because of the variance in metabolite composition of infected and non-infected urine samples.