ChemSpider - building an online database of open spectra
1 like2,495 views
ChemSpider is an online database containing over 30 million chemical compounds and supports the submission and validation of various spectral data, including NMR, IR, Raman, and mass spectra. It leverages community contributions and text-mining techniques to enhance its spectral database, which currently includes over 20,000 spectra. Future plans involve further growth of the database and the exploration of converting spectral figures into accessible spectral data.
ChemSpider - building an online database of open spectra
- 1. ChemSpider – Building an Online Database of Open Spectra Antony J. Williams1,Valery Tkachenko1,Alexey Pshenichnov1, Daniel Lowe2, Carlos Coba3, Kevin Theisen4 and Rudy Potenzone4 1. Royal Society of Chemistry 2. NextMove Software 3. Mestrelab Research 4. iChemLabs LLC Introduction: ChemSpider is an online database of over 30 million chemical compounds from >500 different sources including chemical vendors, online public resources and publications. ChemSpider allows deposition of data including structures, properties, and various forms of spectral data. One activity of the project is to host a searchable database of 1D/2D NMR, IR, Raman and Mass Spectral data. ChemSpider has over 20000 spectra and expands as the community deposits additional data. Sources of Spectral Data: The majority of data are deposited by users of ChemSpider. Submission of spectra in the form of JCAMP-DX, or images/PDF (for all spectra but especially for 2D NMR) are supported. Community-based curators will validate and annotate the data to ensure that only the highest quality data are available on the database. To create a large NMR database we are using “text-mining” to extract spectral data, together with their associated chemical compounds, then simulating visual forms of the spectra,. We have text-mined a large patent corpus to extract many hundreds of thousands of NMR spectra to produce visual depictions as shown in Figure 1. Text mined spectra are of the form: 1H NMR (CDCl3, 400 MHz): δ = 2.57 (m, 4H, Me, C(5a)H), 4.24 (d, 1H, J = 4.8 Hz, C(11b)H), 4.35 (t, 1H, Jb = 10.8 Hz, C(6)H), 4.47 (m, 2H, C(5)H), 4.57 (dd, 1H, J = 2.8 Hz, C(6)H), 6.95 (d, 1H, J = 8.4 Hz, ArH), 7.18–7.94 (m, 11H, ArH) Figure 1: A spectral depiction from converting the text-mined spectrum above. This can be stored in JCAMP to build a spectral database. Spectral Visualization: Spectra are viewed inside the JSpecView spectral display widget1. Zooming, scrolling and integration are possible. 2DNMR spectra are viewed only as images. Figure 2: The JSpecView spectral viewing applet supports viewing JCAMP spectra of 1D NMR, IR, UV-Vis and Mass Spectrometry data. Spectroscopic techniques produce NMR and IR vibrational assignments, and mass fragment peaks. We are now working with iChemLabs HTML5 widgets2 for the display of assignments. Figure 3: Assignments of spectral- structure associations. Selecting the peak at 7.5ppm highlights the protons on the molecule. The assignments are contained in the JCAMP spectral format. Future Directions: We intend to continue to grow the spectral database by encouraging further depositions from the community as well as investigating the possibility of converting spectral figures to spectral data to host in the database. References 1)JSpecView Project: an Open Source Java viewer and converter for JCAMP-DX, and XML spectral data files,http://www.journal.chemistrycentr al.com/content/1/1/31 2)iChemLabs Web Components Spectrum Structure Correlations: http://tinyurl.com/pkz26xf