Fire accelerant classification from GC-MS data of suspected arson cases using machine-learning models.

Using a practical GC-MS dataset containing approximately 4000 suspected arson cases, three machine-learning based classification models were developed and their performances were evaluated. All models trained for classifying the data from fire residue into six categories; no fire accelerants detected or else one of fire accelerants was used within gasoline, kerosene, diesel, solvents, or candle. The classification accuracies of the random forest, supporting vector machine, and convolutional neural network model were 0.88, 0.88, and 0.92, respectively. By calculating feature importance of the random forest model, several potential chemical fingerprints of fire accelerants were discovered.

Tracing the geographical origin of bottled water collected in South Korea via isotopic and elemental analyses.

RATIONALE: The production of bottled water requires a forensic discriminant technique that enables the identification of the brands or accidents caused by intended contaminants. The bottled water poisoning crimes have drawn much attention, and such crimes may recur in the future. The water is colorless and odorless, and thus it is difficult to detect contaminated water through visual observation. Thus, bottled water can be easily exploited for poisoning, and a method for tracing their origin is currently required. METHODS: In this study, a total of 27 brands of bottled water samples were analyzed to determine stable oxygen isotopes, strontium isotopes, major and trace elements. The geographical origin of the water was traced based on the climatic and geographical characteristics of the location from where water was sourced, which was assumed to be reflected in the bottled water. Furthermore, we investigated whether this method can be applied to identify bottled water products. RESULTS: The results demonstrated that the characteristics of the bottled water, including the oxygen stable isotope ratios, reflect the latitude and altitude of bottled water source in South Korea, from the high-latitude region to the coastal regions. In addition, the results indicated that excellent discrimination was achieved using strontium isotopes to identify source areas with different types of bedrock, complex underlying lithologies, and ocean areas in South Korea. A statistical method based on discriminant analysis was applied to measure trace elements, and the results effectively reflected the characteristics of water-rock interactions (cross-validated classification probability: ≥92%). CONCLUSIONS: These data suggest that the geographical characteristics of the source area are well reflected in commercial bottled water in South Korea. The proposed analytical methods can be utilized to trace the geographical origin of different bottled water samples and identify bottled water products used in poisoning crimes.

1 H-NMR with Multivariate Analysis for Automobile Lubricant Comparison.

Identification of suspected automobile-related lubricants could provide valuable information in forensic cases. We examined that automobile lubricants might exhibit the chemometric characteristics to their individual usages. To compare the degree of clustering in the plots, we co-plotted general industrial oils that were highly dissimilar with automobile lubricants in additive compositions. 1 H-NMR spectroscopy was used with multivariate statistics as a tool for grouping, clustering, and identification of automobile lubricants in laboratory conditions. We analyzed automobile lubricants including automobile engine oils, automobile transmission oils, automobile gear oils, and motorcycle oils. In contrast to the general industrial oils, automobile lubricants showed relatively high tendencies of clustering to their usages. Our pilot study demonstrated that the comparison of known and questioned samples to their usages might be possible in forensic fields.

Polymerase recognition of 2-thio-iso-guanine·5-methyl-4-pyrimidinone (iGs·P)--A new DD/AA base pair.

Polymerase specificity is reported for a previously unknown base pair with a non-standard DD/AA hydrogen bonding pattern: 2-thio-iso-guanine·5-methyl-4-pyrimidinone. Our findings suggest that atomic substitution may provide a solution for low fidelity previously associated with enzymatic copying of iso-guanine.

Classification and individualization of used engine oils using elemental composition and discriminant analysis.

The six most common commercial automotive gasoline and diesel engine oils in the Republic of Korea, ZIC A, ZIC XQ RV/SUV, Kixx G1, Kixx RV, and the brand name products HD Premium gasoline and HD Premium diesel, were randomly used in nineteen different vehicles. Samples of seventy-six used engine oils, which were withdrawn from the sumps of those vehicles at different intervals, were analyzed using inductively coupled plasma-optical emission spectrometry (ICP-OES), and statistically compared. Two data analysis strategies were used to interpret and understand the elemental profiles in the multi-dimensional data. Macro (additive elements of Ca, Zn and P) and trace (wear metal elements of Ag, Al, Ba, Cd, Cr, Cu, Fe, Mg, Mo, Na, Ni, Pb and Sn) elements were used as potential markers to determine the brand of oil used and the engine type in which the oil was used, and to trace the individual vehicle for forensic purposes. The discriminant analysis statistical technique was applied, and its prediction ability was assessed. In this study, 92.1%, 82.9% and 92.1% of the cross-validated grouped cases correctly predicted the brand of oil, the engine type and the vehicle that was the source of the oil, respectively.

1H NMR spectroscopic identification of a glue sniffing biomarker.

Organic solvent abuse typically involves sniffing organic solvents to experience the mind-altering conditions they induce. In Republic of Korea, organic solvent abuse is a serious social problem, especially among teenagers. Several studies have addressed the effects of organic solvent abuse on mind and body, but there are no simple methods by which such abuse can be positively identified. In this report, we describe a method for analyzing toluene metabolites (toluene is the main ingredient of glue) in glue-sniffers' urine using (1)H NMR spectroscopy. Toluene is a commonly used solvent in the rubber, paint, plastics, leather, printing, and chemical industries. Inhaled toluene is metabolized to hippuric acid in the liver and excreted in the urine. Hippuric acid is known as a good biomarker for biological monitoring of toluene exposure. We have scanned hippuric acid and other toluene metabolites by NMR spectroscopy and performed statistical multivariate analysis of the data. Based on this analysis, we sought to determine parameters by which glue-sniffing (toluene inhalation) behavior may be verified. We also demonstrate the use of a pattern recognition method for accurate and efficient analysis of NMR data. In comparison to conventional methods, such as mass spectroscopy coupled with liquid chromatography or gas chromatography, nuclear magnetic resonance spectroscopy has several advantages, including simple sample preparation, non-destructive sampling, accuracy, short acquisition time, and reproducibility in the determination of urinary hippuric acid.