BACKGROUND: Osteosarcoma (OS) is the most common primary malignant tumor of bone in children and adolescents. However, few biomarkers of diagnostic significance have been established. In recent years, high-throughput transcriptomic and metabolomic approaches make it possible for studying the levels of thousands of biomarkers simultaneously. METHODS: In this study, we integrated two disparate transcriptomic and metabolomic datasets to find meaningful biomarkers and then used an independent dataset to test the sensibility and specificity of these biomarkers. RESULTS: By using integrated two datasets, we discovered that the biomarkers involved in the glycolysis pathway are highly enriched, including 4 genes (ENO1, TPI1, PKG1 and LDHC) and 2 metabolites (lactate and pyruvate). The 4 genes were significantly down-regulated in OS samples as well as the 2 metabolites. The mixed metabolites + genes signature also outperformed metabolites or genes alone, with recall being 0.813 and F-measure being 0.812. And the AUC value of metabolites + genes classifier was 0.825 (compared to 0.58 for metabolites and 0.821 for genes alone). CONCLUSION: Our findings establish that integrated transcriptomic and metabolomic signature can be used to distinguish OS malignant with good diagnostic accuracy superior to other methods.
BACKGROUND:Osteosarcoma (OS) is the most common primary malignant tumor of bone in children and adolescents. However, few biomarkers of diagnostic significance have been established. In recent years, high-throughput transcriptomic and metabolomic approaches make it possible for studying the levels of thousands of biomarkers simultaneously. METHODS: In this study, we integrated two disparate transcriptomic and metabolomic datasets to find meaningful biomarkers and then used an independent dataset to test the sensibility and specificity of these biomarkers. RESULTS: By using integrated two datasets, we discovered that the biomarkers involved in the glycolysis pathway are highly enriched, including 4 genes (ENO1, TPI1, PKG1 and LDHC) and 2 metabolites (lactate and pyruvate). The 4 genes were significantly down-regulated in OS samples as well as the 2 metabolites. The mixed metabolites + genes signature also outperformed metabolites or genes alone, with recall being 0.813 and F-measure being 0.812. And the AUC value of metabolites + genes classifier was 0.825 (compared to 0.58 for metabolites and 0.821 for genes alone). CONCLUSION: Our findings establish that integrated transcriptomic and metabolomic signature can be used to distinguish OS malignant with good diagnostic accuracy superior to other methods.
Authors: J Toguchida; K Ishizaki; M S Sasaki; Y Nakamura; M Ikenaga; M Kato; M Sugimot; Y Kotoura; T Yamamuro Journal: Nature Date: 1989-03-09 Impact factor: 49.962
Authors: Sha Lou; Benjamin Balluff; Arjen H G Cleven; Judith V M G Bovée; Liam A McDonnell Journal: J Am Soc Mass Spectrom Date: 2016-11-21 Impact factor: 3.109
Authors: Melissa Quintero Escobar; Tássia Brena Barroso Carneiro Costa; Lucas G Martins; Silvia S Costa; André vanHelvoort Lengert; Érica Boldrini; Sandra Regina Morini da Silva; Luiz Fernando Lopes; Daniel Onofre Vidal; Ana C V Krepischi; Mariana Maschietto; Ljubica Tasic Journal: Front Oncol Date: 2020-10-16 Impact factor: 6.244