• Korean Journal of Environmental Biology
• /
• v.39 no.3
• /
• pp.298-310
• /
• 2021

Prediction of the behavior of heavy metals over time is important to evaluate the heavy metal toxicity in algae species. Various modeling studies have been well established, but there is a need for an improved model for predicting the chronic effects of metals on algae species to combine the metal kinetics and biological response of algal cells. In this study, a kinetic dynamics model was developed to predict the copper behavior(5 ㎍ L^-1, 10 ㎍ L^-1, and 15 ㎍ L^-1) for two freshwater algae (Pseudokirchneriella subcapitata and Chlorella vulgaris) in the chronic exposure experiments (8 d and 21 d). In the experimental observations, the rapid change in copper mass between the solutions, extracellular and intracellular sites occurred within initial exposure periods, and then it was slower although the algal density changed with time. Our model showed a good agreement with the measured copper mass in each part for all tested conditions with an elapsed time (R² for P. subcapitata: 0.928, R² for C. vulgaris: 0.943). This study provides a novel kinetic dynamics model that is compromised between practical simplicity and realistic complexity, and it can be used to investigate the chronic effects of heavy metals on the algal population.

• Biomolecules & Therapeutics
• /
• v.19 no.1
• /
• pp.38-44
• /
• 2011

Cisplatin is widely used for various types of cancers. However, its side effects, most notably, renal toxicity often limit its clinical utility. Although previous metabolomic studies reported possible toxicity markers, they used small number of animals and statistical approaches that may not perform best in the presence of intra-group variation. Here, we identified urinary biomarkers associated with renal toxicity induced by cisplatin using NMR-based metabolomics combined with Orthogonal Projections to Latent Structures-Discriminant Analysis (OPLS-DA). Male Sprague-Dawley rats (n=22) were treated with cisplatin (10 mg/kg single dose), and the urines obtained before and after treatment were analyzed by NMR. Multivariable analysis of NMR data presented clear separation between non-treated and treated groups. The OPLS-DA statistical results revealed that 1,3-dimethylurate, taurine, glucose, glycine and branched-chain amino acid (isoleucine, leucine and valine) were significantly elevated in the treated group and that phenylacetylglycine and sarcosine levels were decreased in the treated group. To test the robustness of the approach, we built a prediction model for the toxicity and were able to predict all the unknown samples (n=14) correctly. We believe the proposed NMR-based metabolomics with OPLS-DA approach and the resulting urine markers can be used to augment the currently available blood markers.

• Korean Journal of Environmental Agriculture
• /
• v.12 no.1
• /
• pp.69-80
• /
• 1993

Environmental hazards of a chemical could be assessed by two different approaches : toxicity test and assessment of exposure potentials to human and environmental organisms. For the prediction of environmental fate of chemicals three available computer programs were compared each other and were verified. The results obtained by using these computer programs, PCHEM, EXAMS, and E4CHEM were summarized as follows. The estimated octanol/water partition coefficients by PCHEM were similar to the experimental values in the literature. But the other factors, water solubility and vapor pressure were different from the data in the literature. The simulation results of selected compounds by EXAMS showed similar tendency to the literature results of model field environment. Therefore, this computer program could be utilized to predict the environmental fate of chemicals. E4CHEM program is very simple and this program could predict the ultimate environmental fate of stable chemicals by input of two or three parameters. However, the validity should further be verified in the future field study using more compounds. It is suggested that these approaches could be fully utilized by understanding their limitations to predict the environmental fate of new chemicals under development, to screen the potential environmental pollutants among chemicals already-in use, and to devise measures to minimize the hazards to the environment.

• Microbiology and Biotechnology Letters
• /
• v.27 no.1
• /
• pp.54-61
• /
• 1999

Measurement of inhibition of bioluminescence in Photobacterium phosphoreum has been porposed as a sensitive and rapid procedure to monitor toxic substances. However, at first, $EC_{50}$ which shows degree of toxicity to each toxic substances must be calculated. QSAR (Quantitative Structure Activity Relationship) model can be used to estimate $EC_{50}$ to save time and endeavor. Moderately high correlation coefficients ($r^2{\geq}$ 0.97) were calculated from the linear correlation between $EC_{50}$ and molecular connectivity indices of CAHs (chlorinated aliphatic hydrocarbons)such as $^0X$, $^0X^V$, $^1X$, $^2X$ and $^3X^v_c$ and quadratic correlation between $EC_{50}$ and $^0X$, $^0X^V$, $^2X^V$, $^3X_c$, $^3X^V_c$ and P. It shows that the molecular connection indices in carbon structure is contributed to biological characters with linear relation and that in the other one with quadratic relation. The $EC_{50}$ of chlorophenol derivatives had quadratic relation with the value of octanol/water prtition coefficients ($r^2$=0.99) and linear and quadratic relation with the number of chlorine compound (($r^2{\geq}$0.94). This confirms the already known trend of increasing toxicity with increasing ability of a compound to diffuse through cell membrane and number of chlorine substitution.

• Journal of Korean Society of Environmental Engineers
• /
• v.31 no.2
• /
• pp.125-131
• /
• 2009

Reductive dechlorination of polychlorinated dibenzo-p-dioxins (PCDDs) and its toxicity change were predicted by the linear free energy relationship (LFER) model to assess the zero-valent iron (ZVI) and anaerobic dechlorinating bacteria (ADB) as electron donors in PCDDs dechlorination. Reductive dechlorination of PCDDs involves 256 reactions linking 76 congeners with highly variable toxicities, so is challenging to assess the overall effect of this process on the environmental impact of PCDD contamination. The Gibbs free energies of PCDDs in aqueous solution were updated to density functional theory (DFT) calculation level from thermodynamic results of literatures. All of dechlorination kinetics of PCDDs was evaluated from the linear correlation between the experimental dechlorination kinetics of PCDDs and the calculated thermodynamics of PCDDs. As a result, it was predicted that over 100 years would be taken for the complete dechlorination of octachlorinated dibenzo-p-dioxin (OCDD) to non-chlorinated compound (dibenzo-p-dioxin, DD), and the toxic equivalent quantity (TEQ) of PCDDs could increase to 10 times larger from initial TEQ with the dechlorination process. The results imply that the single reductive dechlorination using ZVI or ADB is not suitable for the treatment strategy of PCDDs contaminated soil, sediment and fly ash. This LFER approach is applicable for the prediction of dechlorination process for organohalogen compounds and for the assessment of electron donating system for treatment strategies.

• Journal of the Korean Society of Safety
• /
• v.31 no.3
• /
• pp.156-161
• /
• 2016

An accident of an ammonia tank pipeline at a storage plant resulted in one death and three injuries in 2014. Many accidents including toxic gas releases and explosions occur in the freezing and refrigerating systems using ammonia. Especially, the consequence can be substantial due to that the large amount of ammonia is usually being used in the refrigeration systems. In this study, offsite consequence analysis has been investigated when ammonia leaks outdoors from large storages. Both flammable and toxic effects are under consideration to calculate the affected area using simulation programs for consequence analysis. ERPG-2 concentration (150 ppm) has been selected to calculate the evacuation distance out of various release scenarios for their dispersions in day or night. For offsite residential, the impact area by flammability is much smaller than that by toxicity. The methodology consists of two steps as followings; 1. Calculation for discharge rates of accidental release scenarios. 2. Dispersion simulation using the discharge rate for different conditions. This proactive prediction for accidental releases of ammonia would help emergency teams act as quick as they can.

• Reproductive and Developmental Biology
• /
• v.32 no.3
• /
• pp.141-146
• /
• 2008

To search a new porcine pheromonal odorant, the models of four type (2D-QSAR, HQSAR, CoMFA & CoMSlA) were derived from quantitative structure-activity relationship (QSAR) between tetrahydrofuran-2-yl family compounds and their observed binding affinity constants (Obs.p$[Od]_{50}$). The optimized CoMFA model (predictability; $r^{2}_{cv.}(q^2)$=0.886 & correlation coefficient: $r^{2}_{ncv.}$=0.984) from ligand based approaches was confirmed as the best model among them. The $N^{1}$-allyl-$N^{2}$-(tetrahydrofuran-2-yl)methyl)oxalamide (P1), 2-(4-trimethylammoniummethylcyclohexyloxy)tetrahydrofurane (P5) and 2-(3-trimethylammoniummethylcyclohexyloxy)tetrahydrofurane (P6) molecules predicted as porcine pheromonal odorant by the CoMFA model were showed relatively high binding affinity constant values (Pred.p$[Od]_{50}=8{\sim}10$) and very lower toxicity values against some sorts of toxicity.

• Journal of Life Science
• /
• v.33 no.6
• /
• pp.490-497
• /
• 2023

Pregnant women may need to take medications to treat preexisting diseases or diseases that develop during pregnancy. However, some drugs may be fetotoxic and lead to, for example, teratogenicity and growth retardation. Predicting the fetotoxicity of drugs is thus important for the health of the mother and fetus. The fetotoxicity of many drugs has not been established because various challenges hinder the ability of researchers to determine their fetotoxicity. The need exists for in silico-based fetotoxicity assessment models, as they can modernize the testing paradigm, improve predictability, and reduce the use of animals and the costs of fetotoxicity testing. In this study, we collected data on the fetotoxicity of drugs and constructed fetotoxicity prediction models based on various machine learning algorithms. We optimized the models for more precise predictions by tuning the hyperparameters. We then performed quantitative performance evaluations. The results indicated that the constructed machine learning-based models had high performance (AUROC >0.85, AUPR >0.9) in fetotoxicity prediction. We also analyzed the feature importance of our model's predictions, which could be leveraged to identify the specific features of drugs that are strongly associated with fetotoxicity. The proposed model can be used to prescreen drugs and drug candidates at a lower cost and in less time. It provides a predictive score for fetotoxicity risk, which may be beneficial in the design of studies on fetotoxicity in human pregnancy.

• Proceedings of the Korea Society of Environmental Toocicology Conference
• /
• 2003.10a
• /
• pp.91-93
• /
• 2003

It has been estimated that the equivalent of approximately $US 50 billion has been spent on research on the behavior and fate of pesticides in the environment since Rachel Carson published “Silent Spring” in 1962. Much of the resulting knowledge has been summarized explicitly in computer algorithms in a variety of empirical, deterministic, and probabilistic simulation models. These models describe and predict the transport, degradation and resultant concentrations of pesticides in various compartments of the environment during and after application. In many cases the known errors of model predictions are large. For this reason they are typically designed to be “conservative”, i.e., err on the side of over-prediction of concentrations in order to err on the side of safety. These predictions are then compared with toxicity data, from tests of the pesticide on a series of standard representative biota, including terrestrial and aquatic indicator species and higher animals (e.g., wildlife and humans). The models' predictions are good enough in some cases to provide screening of those compounds which are very unlikely to do harm, and to indicate those compounds which must be investigated further. If further investigation is indicated a more detailed (and therefore more complicated) model may be employed to give a better estimate, or field experiments may be required. A model may be used to explore “what if” questions leading to possible alternative pesticide usage patterns which give lower potential environmental concentrations and allowable exposures. We are currently at a maturing stage in this research where the knowledge base of pesticide behavior in the environmental is growing more slowly than in the past. However, innovative uses are being made of the explosion in available computer technology to use models to take ever more advantage of the knowledge we have. In this presentation, current developments in the state of the art as practiced in North America and Europe will be presented. Specifically, we will look at the efforts of the ‘Focus’ consortium in the European Union, and the ‘EMWG’ consortium in North America. These groups have been innovative in developing a process and mechanisms for discussion amongst academic, agriculture, industry and regulatory scientists, for consensus adoption of research advances into risk management methodology.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.05a
• /
• pp.276-278
• /
• 2022

All drugs have a specific action in the body, and in many cases, drugs are combinated due to complications or new symptoms during existing drug treatment. In this case, unexpected interactions may occur within the body. Therefore, predicting drug-drug interactions is a very important task for safe drug use. In this study, we propose a deep learning-based predictive model that learns using drug information documents to predict drug interactions that may occur when using multiple drugs. The drug information document was created by combining several properties such as the drug's mechanism of action, toxicity, and target using DrugBank data. And drug information document is pair with another drug documents and used as an input to a deep learning-based predictive model, and the model outputs the interaction between the two drugs. This study can be used to predict future interactions between new drug pairs by analyzing the differences in experimental results according to changes in various conditions.