ポスター発表
- 第2日 5月16日(水) ポスター会場
-
2P-13 PDF
スペクトル相似性ネットワークによる未知化合物スペクトルデータの解析
In mass spectrometry-based small compound analysis, fragment spectra have been widely used to deduce the structures of analyte compounds. However, gaining insight into the complex chemical nature of matrices containing a large number of “unknown unknown” compounds, whose structures have not been described, is still quite difficult to accomplish.
Fragment similarity network, also called as MS/MS spectral network or molecular networking, is a method to assess and visualize the similarities of fragment spectra of structurally related compounds. We have developed a data analysis framework employing fragment similarity network to organize, classify, (sub)structurally annotate and visualize the complex fragment spectral dataset of unknown compounds. Using reference spectral dataset and in-silico fragments generated from compound structure database, multi-layer representation of spectral similarity networks effectively visualizes and categorizes the spectra of unknown compounds in sample. Structural information systematically extracted from the compounds in reference and in-silico spectral networks provides structural insights into the unknown compounds. This versatile method can be applied to spectral dataset containing a wide range of small molecules. We demonstrate the application of this method using case studies.