• Mass Spectrometry Letters
• /
• v.10 no.3
• /
• pp.93-101
• /
• 2019

Algorithms and software for predicting tandem mass spectra have been developed in recent years. In this work, we explore how distinct in silico $MS^2$ spectra are predicted for isomers, i.e. compounds having the same formula and similar molecular structures, to differentiate between them. We used the CFM-ID 2.0/3.0 predictor with regard to (a) test compounds, whose experimental mass spectra had been randomly sampled from the MassBank of North America (MoNA) collection, and to (b) the most widespread isomers of test compounds searched in the PubChem database. In the first validation test, in silico mass spectra constitute a reference library, and library searches are performed for test experimental spectra of "unknowns". The searches led to the true positive rate (TPR) of ($46-48{\pm}10$)%. In the second test, in silico and experimental spectra were interchanged and this resulted in a TPR of ($58{\pm}10$)%. There were no significant differences between results obtained with different metrics of spectral similarity and predictor versions. In a comparison of test compounds vs. their isomers, a statistically significant correlation between mass spectral data and structural features was observed. The TPR values obtained should be regarded as reasonable results for predicting tandem mass spectra of related chemical structures.

• Journal of Korean Institute of Industrial Engineers
• /
• v.37 no.4
• /
• pp.400-407
• /
• 2011

Modeling and analysis of high-dimensional spectral data provide an opportunity to uncover inherent patterns in various information-rich data. Orthogonal signal correction (OSC) a preprocessing technique has been widely used to remove unwanted variations of spectral data that do not contribute to prediction or classification. In the present study we propose a novel OSC algorithm called adjusted direct OSC to improve visualization and the ability of classification. Experimental results with real mass spectral data from condom lubricants demonstrate the effectiveness of the proposed approach.

• Publications of The Korean Astronomical Society
• /
• v.32 no.1
• /
• pp.73-75
• /
• 2017

Debris disks are important observational clues to understanding on-going planetary system formation. They are usually identified by significant mid-infrared excess on top of the photospheric emission of a central star on the basis of prediction from J-, H-, and Ks-band fluxes and the stellar model spectra. For bright stars, 2MASS near-infrared fluxes suffer large uncertainties due to the near-infrared camera saturation. Therefore we have performed follow-up observations with the IRSF 1.4 m near-infrared telescope located in South Africa to obtain accurate J-, H-, and Ks-band fluxes of the central stars. Among 754 main-sequence stars which are detected in the AKARI $18{\mu}m$ band, we have performed photometry for 325 stars with IRSF. As a result, we have successfully improved the flux accuracy of the central stars from 9.2 % to 0.5 % on average. Using this dataset, we have detected $18{\mu}m$ excess emission from 57 stars in our samples with a $3{\sigma}$ level. We find that some of them have high ratios of the excess to the photospheric emission even around very old stars, which cannot be explained by the current planet-formation theories.

• Korean Journal of Breeding Science
• /
• v.42 no.5
• /
• pp.448-454
• /
• 2010

The objective of this study was to rapidly evaluate fatty acids in a collection of foxtail millet (Setaria italica (L.) P. Beauv) of different origins so that this information could be disseminated to breeders to advance germplasm use and breeding. To develop the calibration equations for rapid and nondestructive evaluation of fatty acid content, near-infrared reflectance spectroscopy (NIRs) spectra (1104-2494 nm) of samples ground into flour (n=100) were obtained using a dispersive spectrometer. A modified partial least-squares model was developed to predict each component. For foxtail millet germplasm, our models returned coefficients of determination ($R^2$) of 0.91, 0.89, 0.98 and 0.98 for strearic acid, oleic acid, linoleic acid, and total fatty acids, respectively. The prediction of the external validation set (n=10) showed significant correlation between references values and NIRs values ($r^2=0.97$, 0.91, 0.99 for oleic, linoleic, and total fatty acids, respectively). Standard deviation/standard error of cross-validation (SD/SECV) values were greater than 3 (3.11, 5.45, and 7.50 for oleic, linoleic, and total fatty acids, respectively). These results indicate that these NIRs equations are functional for the mass screening and rapid quantification of the oleic, linolenic, and total fatty acids characterizing foxtail millet germplasm. Among the samples, IT153491 showed an especially high content of fatty acids ($84.06mg\;g^{-1}$), whereas IT188096 had a very low content ($29.92mg\;g^{-1}$).

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.57 no.1
• /
• pp.29-34
• /
• 2012

The objective of this study was to rapidly evaluate fatty acids in a collection of millet (Panicum miliaceum subsp. miliaceum) of different origins so that this information could be disseminated to breeders to advance germplasm use and breeding. To develop the calibration equations for rapid and nondestructive evaluation of fatty acid content, near-infrared reflectance spectroscopy (NIRs) spectra (1104-2494 nm) of samples ground into flour ($n$=100) were obtained using a dispersive spectrometer. A modified partial least-squares model was developed to predict each component. For foxtail millet germplasm, our models returned coefficients of determination ($R^2$) of 0.89, 0.89, 0.89, and 0.92 for palmitic acid, oleic acid, linoleic acid, and total fatty acids, respectively. The prediction of the external validation set (n=10) showed significant correlation between references values and NIRs values ($r^2$=0.64, 0.90, 0.79, and 0.89 for palmitic acid, oleic acid, linoleic acid, and total fatty acids, respectively). Standard deviation/standard errors of cross-validation (SD/SECV) values were close to 3 (2.62, 2.40, 1.85, and 2.23 for palmitic acid, oleic acid, linoleic acid, and total fatty acids, respectively). These results indicate that these NIRs equations are functional for the mass screening and rapid quantification of the oleic and total fatty acids characterizing millet germplasm. Among the samples, IT153514 showed an especially high content of fatty acids ($48.14mg\;g^{-1}$), whereas IT123909 had a very low content ($34.44mg\;g^{-1}$).

• The Bulletin of The Korean Astronomical Society
• /
• v.38 no.2
• /
• pp.38.2-38.2
• /
• 2013

Several studies have reported the presence of sodium excess objects having neutral atomic absorption lines at $5895{\AA}$ (NaD) and $8190{\AA}$ that are deeper than expected based on stellar population models that match the stellar continuum. The origin of these lines is therefore hotly debated. van Dokkum & Conroy proposed that low-mass stars (0.3M) are more prevalent in massive early-type galaxies, which may lead to a strong NaI 8190 line strength. It is necessary to test this prediction, however, against other prominent optical line indices such as NaD, Mgb, and Fe5270, which can be measured with a significantly higher signal-to-noise ratio than NaI 8190. We identified a new sample of roughly one thousand NaD excess objects (NEOs; ~8% of galaxies in the sample) based on NaD line strength in the redshift range 0.00${\alpha}$-enhanced" ([${\alpha}/Fe$] ~ 0.3), "metal-rich" ([Z/H] ~ 0.3), and, especially, "Na-enhanced" ([Na/Fe] ~ 0.3).

• The Bulletin of The Korean Astronomical Society
• /
• v.39 no.1
• /
• pp.47.2-47.2
• /
• 2014

Several studies have reported the presence of sodium excess objects having neutral atomic absorption lines at $5895{\AA}$ (NaD) and $8190{\AA}$ that are deeper than expected based on stellar population models that match the stellar continuum. The origin of these lines is therefore hotly debated. van Dokkum & Conroy proposed that low-mass stars (0.3M) are more prevalent in massive early-type galaxies, which may lead to a strong NaI 8190 line strength. It is necessary to test this prediction, however, against other prominent optical line indices such as NaD, Mgb, and Fe5270, which can be measured with a significantly higher signal-to-noise ratio than NaI 8190. We identified a new sample of roughly one thousand NaD excess objects (NEOs; ~8% of galaxies in the sample) based on NaD line strength in the redshift range 0.00$H{\beta}$ line strengths and significant emission lines, which are indicative of the presence of young stellar populations. This result implies that the presence of the interstellar medium and/or dust contributes to the increase in NaD line strengths observed for these galaxies.