• Korean Chemical Engineering Research
• /
• 제54권5호
• /
• pp.621-629
• /
• 2016

EU의 REACH 제도 도입에 따라 각종 화학물질에 대한 독성 및 활성 정보 확보를 위해 화학물질의 분자구조 정보를 기반으로 화학물질의 독성 및 활성을 예측하는 정량적구조활성관계(QSAR)에 대한 연구가 최근 활발히 진행되고 있다. QSAR 모델에 사용되는 분자표현자는 매우 다양하기 때문에 화학물질의 물성 및 활성을 잘 표현할 수 있는 주요한 분자표현자를 선택하는 과정은 QSAR 모델 개발에 있어 중요한 부분이다. 본 연구에서는 화학물질의 분자구조 정보를 나타내는 주요 분자표현자의 통계적 선택 방법과 부분최소자승법(Partial least square: PLS) 기반의 새로운 QSAR 모델을 제안하였다. 제안된 QSAR 모델은 130종의 폴리염화바이페닐(Polychlorinated biphenyl: PCB)에 대한 분배계수(log P)와 14종의 PCBs에 대한 반수 치사 농도(Lethal concentration 50%: $LC_{50}$) 예측에 사용되고, 제안된 QSAR 모델 예측 정확도는 기존의 OECD QSAR Toolbox에서 제공하는 QSAR 모델과 비교하였다. 관심 화학물질의 분자표현자와 활성정보 간의 높은 상관관계를 갖는 주요 분자표현자를 선별하기 위해서, 상관계수(r)와 variable importance on projections (VIP)기법을 적용하였으며, 화학물질의 독성 및 활성정보를 예측하기 위해 선별된 분자표현자와 활성정보를 이용해 부분최소자승법(PLS)를 사용하였다. 회귀계수($R^2$)와 prediction residual error sum of square (PRESS)을 이용한 성능평가결과, 제안된 QSAR 모델은 OECD QSAR Toolbox의 QSAR 모델보다 PCBs의 log P와 $LC_{50}$에 대하여 각각 26%, 91% 향상된 예측력을 나타내었다. 본 연구에서 제안된 계산독성학 기반의 QSAR 모델은 화학물질의 독성 및 활성정보에 대한 예측력을 향상시킬 수 있고 이러한 방법은 유독 화학물질의 인체 및 환경 위해성 평가에 기여할 것으로 판단된다.

• 분석과학
• /
• 제24권6호
• /
• pp.544-555
• /
• 2011

본 연구는 농약류에 대하여 구조-활성의 정량적 관계(QSAR)를 이용하여 무지개 송어(학명: Oncorhynchus mykiss)의 급성 독성을 예측-분석하는 과정을 수행하였다. 모델 구현을 위해 사용된 275종의 농약류에 대한 수중 독성(96h $LC_{50}$) 값은 DEMETRA프로젝트의 데이터를 사용하였다. 예측 모델에 사용된 2차원 분자 표현자는 PreADMET프로그램으로부터 계산을 하였고, 선형 (다중 선형 회귀 방법)모델과 비선형(서포트 벡터 머신, 인공 신경망) 학습 방법들은 실험값과 예측값의 적합도를 고려하여 최적화 되었다. 데이터 전처리 과정을 거친 뒤에, 5묶음 교차 검증과정을 포함한 모집단 기반 전진 선택법을 통해서 각 학습 방법의 최적의 표현자 집합을 결정하였다. 가장 좋은 결과는 SVM 방법 ($R^2_{CV}$=0.677, RMSECV=0.887, MSECV=0.674) 이었고, EU의 규제 기준에 따른 분류에서는 87%의 정확도를 나타내었다. MLR방법을 통해서는 무지개 송어의 급성 독성에 대하여 독성을 나타내는 농약류의 구조적 특징과 지질 층과의 상호작용을 설명할 수 있었다. 개발된 모든 모델들은 5묶음 교차 검증과 Y-scrambling test을 통해 검증되었다.

• Bulletin of the Korean Chemical Society
• /
• 제35권12호
• /
• pp.3637-3641
• /
• 2014

In this study, we analyzed the toxicity of mixtures of dimethylformamide (DMF) and methyl ethyl ketone (MEK) or DMF and toluene (TOL) and predicted their toxicity using quantitative structure-activity relationships (QSAR). A QSAR model for single substances and mixtures was analyzed using multiple linear regression (MLR) by taking into account the statistical parameters between the observed and predicted $EC_{50}$. After preprocessing, the best subsets of descriptors in the learning methods were determined using a 5-fold cross-validation method. Significant differences in physico-chemical properties such as boiling point (BP), specific gravity (SG), Reid vapor pressure (rVP), flash point (FP), low explosion limit (LEL), and octanol/water partition coefficient (Pow) were observed between the single substances and the mixtures. The $EC_{50}$ of the mixture of DMF and TOL was significantly lower than that of DMF. The mixture toxicity was directly related to the mixing ratio of TOL and MEK (MLR $EC_{50}$ equation = $1.76997-1.12249{\times}TOL+1.21045{\times}MEK$), as well as to SG, VP, and LEL (MLR equation $EC_{50}=15.44388-19.84549{\times}SG+0.05091{\times}VP+1.85846{\times}LEL$). These results show that QSAR-based models can be used to quantitatively predict the toxicity of mixtures used in manufacturing industries.

• 대한안전경영과학회지
• /
• 제26권3호
• /
• pp.59-70
• /
• 2024

This study aims to estimate the scope of damage impact with a real-life explosion case and a damage prediction program (ALOHA) and suggest measures to reduce risk by comparing and analyzing the results using a Probit model. After applying it to the ALOHA program, the toxicity, overpressure, and radiant heat damage of 5 tons of storage scopes between 66 to 413 meters, and the real-life case also demonstrated that most of the damage took place within 300 meters of the LPG gas station. In the Probit analysis, the damages due to radiant heat were estimated as first-degree burns (13-50%), while structural damage (0-75%) and glass window breakage (94-100%) were expected from overpressure, depending on the storage volume. After comparing the real-life case and the damage prediction program, this study concluded that the ALOHA program could be used as the scope of damage impacts is nearly the same as the actual case; it also concluded that the analysis using the Probit model could reduce risks by applying calculated results and predicting the probability of human casualties and structural damages.

• Toxicological Research
• /
• 제32권4호
• /
• pp.289-300
• /
• 2016

Non-genotoxic carcinogens are substances that induce tumorigenesis by non-mutagenic mechanisms and long term rodent bioassays are required to identify them. Recent studies have shown that transcription profiling can be applied to develop early identifiers for long term phenotypes. In this study, we used rat liver expression profiles from the NTP (National Toxicology Program, Research Triangle Park, USA) DrugMatrix Database to construct a gene classifier that can distinguish between non-genotoxic carcinogens and other chemicals. The model was based on short term exposure assays (3 days) and the training was limited to oxidative stressors, peroxisome proliferators and hormone modulators. Validation of the predictor was performed on independent toxicogenomic data (TG-GATEs, Toxicogenomics Project-Genomics Assisted Toxicity Evaluation System, Osaka, Japan). To build our model we performed Random Forests together with a recursive elimination algorithm (VarSelRF). Gene set enrichment analysis was employed for functional interpretation. A total of 770 microarrays comprising 96 different compounds were analyzed and a predictor of 54 genes was built. Prediction accuracy was 0.85 in the training set, 0.87 in the test set and increased with increasing concentration in the validation set: 0.6 at low dose, 0.7 at medium doses and 0.81 at high doses. Pathway analysis revealed gene prominence of cellular respiration, energy production and lipoprotein metabolism. The biggest target of toxicogenomics is accurately predict the toxicity of unknown drugs. In this analysis, we presented a classifier that can predict non-genotoxic carcinogenicity by using short term exposure assays. In this approach, dose level is critical when evaluating chemicals at early time points.

• 한국컴퓨터정보학회논문지
• /
• 제21권8호
• /
• pp.57-64
• /
• 2016

Initial prevention activities and rapid propagation conditions is the most important to prevent diffusion of water pollution. If water pollutants flow into streams river or main stresm located in environmental conservation area or water intake facilities, we must predict immediately arrival time and the diffusion concentration to the proactive. National Institute of Environmental Research developed water pollution incident response prediction system linking dam and movable weir. the system is mathematical model which is updated daily. Therefore it can quickly predict the arrival time and the diffusion concentration when there are accident of oil spills and hazardous chemicals. Also we equipped with mathematical model and toxicity model of EFDC(Environmental Fluid Dynamics Code) to calculate the arrival time and the diffusion concentration. However these systems offer the services of an offline manner than real-time control services. we have ensured the reliability of data collection and have developed a real-time water quality measurement data transmission device by using the data linkage utilizing a mode bus communication and a commercial SCADA system, in particular, we implemented to be able to do real-time water quality prediction through information infrastructure of the water quality integrated management business created by utilizing the construction of the real-time prediction system that utilizes the data collected, the Open map, the visual representation using charts API and development of integrated management system development based on web maps.

• 한국환경성돌연변이발암원학회지
• /
• 제21권2호
• /
• pp.118-121
• /
• 2001

The predictive screening of various molecular descriptors for predicting carcinogenic, mutagenic, teratogenic and alkylation activity of chlorinated disinfection byproducts (DBPs) has been investigated for the application of quantitative structure-activity relationships (QSAR). The toxicity index for 29 compounds were computed by the PASS program and active values were employed in this study. Studies show that different descriptors account for the model equation of each genotoxic endpoint and that thermodynamic descriptors significantly played a major role on prediction of endpoints of chlorinated aliphatic compounds.

• 한국지하수토양환경학회지:지하수토양환경
• /
• 제22권5호
• /
• pp.30-39
• /
• 2017

In this study, the predictive toxicity of barley Hordeum vulgare was estimated using a modified terrestrial biotic ligand model (TBLM) to account for the toxic effects of $CuOH^+$ and $CuCO_3(aq)$ generated at pH 7 or higher, and this was compared to that from the original TBLM. At pH values higher than 7, the difference in $EA_{50}\{Cu^{2+}\}$ (half maximal effective activity of $Cu^{2+}$) between the two models increased with increasing pH. As Mg concentration increased from 8.24 to 148 mg/L in the pH range of 5.5 to 8.5, the difference in $EA_{50}\{Cu^{2+}\}$ increased, and it reached its maximum at pH 8. The difference in $EC_{50}[Cu]_T$ (half maximal effective concentration of Cu) between the two models increased as dissolved organic carbon (DOC) concentration increased when pH was above 7. Thus, for soils with alkaline pH, the toxic effect of $CuOH^+$ and $CuCO_3(aq)$ are greater at higher salt and DOC concentrations. The acceptable Cu concentration in soil porewater can be estimated by the modified TBLM through deterministic method at pH levels higher than 7, while combination of TBLM and species sensitivity distribution through the probabilistic method could be utilized at pH levels lower than 7.

• Toxicological Research
• /
• 제30권1호
• /
• pp.1-5
• /
• 2014

Xenograft models of human cancer play an important role in the screening and evaluation of candidates for new anticancer agents. The models, which are derived from human tumor cell lines and are classified according to the transplant site, such as ectopic xenograft and orthotopic xenograft, are still utilized to evaluate therapeutic efficacy and toxicity. The metastasis model is modified for the evaluation and prediction of cancer progression. Recently, animal models are made from patient-derived tumor tissue. The patient-derived tumor xenograft models with physiological characters similar to those of patients have been established for personalized medicine. In the discovery of anticancer drugs, standard animal models save time and money and provide evidence to support clinical trials. The current strategy for using xenograft models as an informative tool is introduced.

• 대한화학회지
• /
• 제53권6호
• /
• pp.653-662
• /
• 2009

흡수, 분포, 대사, 배설 특성 및 독성을 예측하기 위한 효과적인 툴을 개발하는 것은 신약개발의 초기단계에서 NCE(new chemical entity)에 대한 가장 중요한 업무 중의 하나이다. 최근에 이런 시도중의 하나로서 ADME/T(absorption, distribution, metabolism, excretion, toxicity)관련 성질들의 예측에 support vector machine(SVM)을 이용하고 있다. 그리고 SVM은 ADME/T 성질들을 정확하게 예측하는데 많이 사용 되고 있다. 그러나 SVM 모델링에 두 가지 문제가 있다. 특성 선택(feature selection) 과 매개변수 설정(parameter setting)은 여전히 해결해야 할 과제이다. 이 두 가지 문제들은 SVM 분류의 효율성과 정확도에 결정적인 영향을 끼친다. 특히 특성 선택과 최적화된 SVM 변수의 설정은 서로 영향을 주기 때문에 동시에 다루어져야 한다. 여기서 우리는 genetic algorithm(GA) – 특성 선택에 사용 – 과 grid search(GS) method– 변수최적화에 사용 – 두 가지를 통합하는 효과적인 해결책을 제시하였다. ADME/T관련 성질 중 하나인 심장부정맥을 야기시키는 hERG 이온채널 저해제 분류 모델이 여기서 제안된 GA-GS-SVM을 위해 할당되고 테스트 되었다. 1891개의 화합물을 가지는 트레이닝 셋으로 단일 모델 3개, 앙상블 모델 3개, 총 6개의 모델을 만들었고 175개의 외부 데이터를 테스트 셋으로 사용하여 검증하였다. 데이터의 불균형 문제를 해결하기 위하여 GA-GS-SVM 단일 모델에 의한 예측 정확도와 GA-GS-SVM 앙상블 모델 예측 정확도를 비교하였으며, 앙상블모델을 사용하여 예측의 정확도를 높일 수 있었다.