Advancing Fifth Percentile Hazard Concentration Estimation Using Toxicity-Normalized Species Sensitivity Distributions.

The species sensitivity distribution (SSD) is an internationally accepted approach to hazard estimation using the probability distribution of toxicity values that is representative of the sensitivity of a group of species to a chemical. Application of SSDs in ecological risk assessment has been limited by insufficient taxonomic diversity of species to estimate a statistically robust fifth percentile hazard concentration (HC5). We used the toxicity-normalized SSD (SSDn) approach, (Lambert, F. N.; Raimondo, S.; Barron, M. G. Environ. Sci. Technol.2022,56, 8278-8289), modified to include all possible normalizing species, to estimate HC5 values for acute toxicity data for groups of carbamate and organophosphorous insecticides. We computed mean and variance of single chemical HC5 values for each chemical using leave-one-out (LOO) variance estimation and compared them to SSDn and conventionally estimated HC5 values. SSDn-estimated HC5 values showed low uncertainty and high accuracy compared to single-chemical SSDs when including all possible combinations of normalizing species within the chemical-taxa grouping (carbamate-all species, carbamate-fish, organophosphate-fish, and organophosphate-invertebrate). The SSDn approach is recommended for estimating HC5 values for compounds with insufficient species diversity for HC5 computation or high uncertainty in estimated single-chemical HC5 values. Furthermore, the LOO variance approach provides SSD practitioners with a simple computational method to estimate confidence intervals around an HC5 estimate that is nearly identical to the conventionally estimated HC5.

Assessment of a New Approach Method for Grouped Chemical Hazard Estimation: The Toxicity-Normalized Species Sensitivity Distribution (SSDn).

New approach methods are being developed to address the challenges of reducing animal testing and assessing risks to the diversity of species in aquatic environments for the multitude of chemicals with minimal toxicity data. The toxicity-normalized species sensitivity distribution (SSDn) approach is a novel method for developing compound-specific hazard concentrations using data for toxicologically similar chemicals. This approach first develops an SSDn composed of acute toxicity values for multiple related chemicals that have been normalized by the sensitivity of a common species tested with each compound. A toxicity-normalized hazard concentration (HC5n) is then computed from the fifth percentile of the SSDn. Chemical-specific HC5 values are determined by back-calculating the HC5n using the chemical-specific sensitivity of the normalization species. A comparison of the SSDn approach with the single-chemical SSD method was conducted by using data for nine transition metals to generate and compare HC5 values between the two methods. We identified several guiding principles for this method that, when applied, resulted in accurate HC5 values based on comparisons with results from single-metal SSDs. The SSDn approach shows promise for developing statistically robust hazard concentrations when adequate taxonomic representation is not available for a single chemical.

Relationships Between Aquatic Toxicity, Chemical Hydrophobicity, and Mode of Action: Log Kow Revisited.

Relationships between toxicity and chemical hydrophobicity have been known for nearly 100 years in mammals and fish, typically using the log of the octanol:water partition coefficient (Kow). The current study reassessed the influence of mode of action (MOA) on acute aquatic toxicity-log Kow relationships using a comprehensive database of 617 organic chemicals with curated and standardized acute toxicity data that did not exceed solubility limits, their consensus log Kow values, and weight of evidence-based MOA classifications (including 6 broad and 26 specific MOAs). A total of 166 significant (p < 0.05) log Kow-toxicity models were developed across six taxa groups that included QSARs for 5 of the broad and 13 of the specific MOAs. In this study, we demonstrate that QSARs based on MOAs can significantly increase LC50 prediction accuracy for specific acting chemicals. Prediction accuracy increases when QSARs are built based on highly specific MOAs, rather than broad MOA classifications. Additionally, we demonstrate that building QSAR models with chemicals in specific MOA groupings, rather than broader MOA groups leads to significantly better estimates. We also evaluated the differences between models developed from mass-based (µg/L) and mole-based (µmol/L) toxicity data and demonstrate that both are suitable for QSAR development with no clear trend in greater model accuracy. Overall, the results reveal that, despite high variance in all taxa and MOA groups, specific MOA-based models can improve the accuracy of aquatic toxicity predictions over more general groupings.Please check and confirm that the authors and their respective affiliations have been correctly identified and amend if necessary.The affiliations are correct.

Long-Term Ecological Impacts from Oil Spills: Comparison of Exxon Valdez, Hebei Spirit, and Deepwater Horizon.

The long-term ecological impacts of the Exxon Valdez oil spill (EVOS) are compared to two extensively studied and more recent large spills: Deepwater Horizon (DWH) and the Hebei Spirit oil spill (HSOS). Each of the three spills differed in magnitude and duration of oil released, environmental conditions, ecological communities, response and clean up measures, and ecological recovery. The EVOS began on March 24, 1989, and released 40.8 million liters of Alaska North Slope crude oil into the cold, nearly pristine environment of Prince William Sound, Alaska. EVOS oiled wildlife and rocky intertidal shorelines and exposed early life stages of fish to embryotoxic levels of polycyclic aromatic hydrocarbons (PAH). Long-term impacts following EVOS were observed on seabirds, sea otters, killer whales, and subtidal communities. The DWH spill began on April 20, 2010, and released 507 million liters of light Louisiana crude oil from 1600 m on the ocean floor into the Gulf of Mexico over an 87-day period. The DWH spill exposed a diversity of complex aquatic communities in the deep ocean, offshore pelagic areas, and coastal environments to petroleum hydrocarbons. Large-scale persistent ecological effects included impacts to deep ocean corals, failed recruitment of oysters over multiple years, damage to coastal wetlands, and reduced dolphin, sea turtle, and seabird populations. The HSOS began on December 7, 2007, and released approximately 13 million liters of Middle East crude oils into ecologically sensitive areas of the Taean area of western Korea. Environmental conditions and the extensive initial cleanup of HSOS oil stranded on shorelines limited the long-term impacts to changes in composition and abundance of intertidal benthic communities. Comparisons of EVOS, DWH, and HSOS show the importance and complexity of the interactions among the environment, oil spill dynamics, affected ecological systems, and response actions.

A Bayesian network approach to refining ecological risk assessments: Mercury and the Florida panther (Puma concolor coryi).

Traditionally hazard quotients (HQs) have been computed for ecological risk assessment, often without quantifying the underlying uncertainties in the risk estimate. We demonstrate a Bayesian network approach to quantitatively assess uncertainties in HQs using a retrospective case study of dietary mercury (Hg) risks to Florida panthers (Puma concolor coryi). The Bayesian network was parameterized, using exposure data from a previous Monte Carlo-based assessment of Hg risks (Barron et al., 2004. ECOTOX 13:223), as a representative example of the uncertainty and complexity in HQ calculations. Mercury HQs and risks to Florida panthers determined from a Bayesian network analysis were nearly identical to those determined using the prior Monte Carlo probabilistic assessment and demonstrated the ability of the Bayesian network to replicate conventional HQ-based approaches. Sensitivity analysis of the Bayesian network showed greatest influence on risk estimates from daily ingested dose by panthers and mercury levels in prey, and less influence from toxicity reference values. Diagnostic inference was used in a high-risk scenario to demonstrate the capabilities of Bayesian networks for examining probable causes for observed effects. Application of Bayesian networks in the computation of HQs provides a transparent and quantitative analysis of uncertainty in risks.

Mixed phylogenetic signal in fish toxicity data across chemical classes.

Chemical use in society is growing rapidly and is one of the five major pressures on biodiversity worldwide. Since empirical toxicity studies of pollutants generally focus on a handful of model organisms, reliable approaches are needed to assess sensitivity to chemicals across the wide variety of species in the environment. Phylogenetic comparative methods (PCM) offer a promising approach for toxicity extrapolation incorporating known evolutionary relationships among species. If phylogenetic signal in toxicity data is high, i.e., closely related species are more similarly sensitive as compared to distantly related species, PCM could ultimately help predict species sensitivity when toxicity data are lacking. Here, we present the largest ever test of phylogenetic signal in toxicity data by combining phylogenetic data from fish with acute mortality data for 42 chemicals spanning 10 different chemical classes. Phylogenetic signal is high for some chemicals, particularly organophosphate pesticides, but not necessarily for many chemicals in other classes (e.g., metals, organochlorines). These results demonstrate that PCM may be useful for toxicity extrapolation in untested species for those chemicals with clear phylogenetic signal. This study provides a framework for using PCM to understand the patterns and causes of variation in species sensitivity to pollutants.

Calcification in Caribbean reef-building corals at high pCO2 levels in a recirculating ocean acidification exposure system.

Projected increases in ocean pCO2 levels are anticipated to affect calcifying organisms more rapidly and to a greater extent than other marine organisms. The effects of ocean acidification (OA) have been documented in numerous species of corals in laboratory studies, largely tested using flow-through exposure systems. We developed a recirculating ocean acidification exposure system that allows precise pCO2 control using a combination of off-gassing measures including aeration, water retention devices, venturi injectors, and CO2 scrubbing. We evaluated the recirculating system performance in off-gassing effectiveness and maintenance of target pCO2 levels over an 84-day experiment. The system was used to identify changes in calcification and tissue growth in response to elevated pCO2 (1000 µatm) in three reef-building corals of the Caribbean: Pseudodiploria clivosa, Montastraea cavernosa, and Orbicella faveolata. All three species displayed an overall increase in net calcification over the 84-day exposure period regardless of pCO2 level (control +0.28- 1.12 g, elevated pCO2 +0.18- 1.16 g), and the system was effective at both off-gassing acidified water to ambient pCO2 levels, and maintaining target elevated pCO2 levels over the 3-month experiment.

Photoenhanced Toxicity of Weathered Crude Oil in Sediment and Water to Larval Zebrafish.

Solar radiation exposure can increase the toxicity of bioaccumulated oil compounds in a diversity of aquatic species. We investigated the photoenhanced toxicity of weathered South Louisiana crude oil in sediment and water accommodated fractions (WAF) to larval zebrafish. Larvae were first exposed for 24 h to one of six treatments: no oil (sediment or water), 7.5 g oil/kg sediment, oil-only WAF, oil WAF plus the dispersant Corexit 9500A, or dispersant alone. Larvae were then exposed to high or low levels of sunlight in control water for 3 or 3.5 h. Hydrocarbon concentrations were measured in exposure media, including alkanes, polycyclic aromatic compounds and total petroleum hydrocarbons. Significant phototoxicity was observed in larvae exposed to oiled sediment, oil-only WAF, and oil plus dispersant WAF. The results indicated that petroleum from the northern Gulf of Mexico can be phototoxic to larval fish exposed to oil in either the water column or sediment.

Framework for Optimizing Selection of Interspecies Correlation Estimation Models to Address Species Diversity and Toxicity Gaps in an Aquatic Database.

The Chemical Aquatic Fate and Effects (CAFE) database is a tool that facilitates assessments of accidental chemical releases into aquatic environments. CAFE contains aquatic toxicity data used in the development of species sensitivity distributions (SSDs) and the estimation of hazard concentrations (HCs). For many chemicals, gaps in species diversity and toxicity data limit the development of SSDs, which may be filled with Interspecies Correlation Estimation (ICE) models. Optimization of ICE model selection and integration ICE-predicted values into CAFE required a multistep process that involved the use of different types of data to assess their influence on SSDs and HC estimates. Results from multiple analyses showed that SSDs supplemented with ICE-predicted values generally produced HC5 estimates that were within a 3-fold difference of estimates from measured SSDs (58%-82% of comparisons), but that were often more conservative (63%-76% of comparisons) and had lower uncertainty (90% of comparisons). ICE SSDs did not substantially underpredict toxicity (<10% of comparisons) when compared to estimates from measured SSD. The incorporation of ICE-predicted values into CAFE allowed the development of >800 new SSDs, increased diversity in SSDs by an average of 34 species, and augmented data for priority chemicals involved in accidental chemical releases.

Photoenhanced Toxicity of Petroleum to Aquatic Invertebrates and Fish.

Photoenhanced toxicity is a distinct mechanism of petroleum toxicity that is mediated by the interaction of solar radiation with specific polycyclic aromatic compounds in oil. Phototoxicity is observed as a twofold to greater than 1000-fold increase in chemical toxicity to aquatic organisms that also have been exposed to light sources containing sufficient quantity and quality of ultraviolet radiation (UV). When tested under natural sunlight or laboratory sources of UV, fresh, and weathered middle distillates, crudes and heavy oils can exhibit photoenhanced toxicity. These same products do not exhibit phototoxicity in standard test protocols because of low UV irradiance in laboratory lighting. Fresh, estuarine, and marine waters have been shown to have sufficient solar radiation exposure to elicit photoenhanced toxicity, and a diversity of aquatic invertebrate and fish species can exhibit photoenhanced toxicity when exposed to combinations of oil and UV. Risks of photoenhanced toxicity will be greatest to early life stages of aquatic organisms that are translucent to UV and that inhabit the photic zone of the water column and intertidal areas exposed to oil.

Spatially explicit assessment of estuarine fish after Deepwater Horizon oil spill: trade-off in complexity and parsimony.

Evaluating long-term contaminant effects on wildlife populations depends on spatial information about habitat quality, heterogeneity in contaminant exposure, and sensitivities and distributions of species integrated into a systems modeling approach. Rarely is this information readily available, making it difficult to determine the applicability of realistic models to quantify population-level risks. To evaluate the trade-offs between data demands and increased specificity of spatially explicit models for population-level risk assessments, we developed a model for a standard toxicity test species, the sheepshead minnow (Cyprinodon variegatus), exposed to oil contamination following the Deepwater Horizon oil spill and compared the output with various levels of model complexity to a standard risk quotient approach. The model uses habitat and fish occupancy data collected over five sampling periods throughout 2008-2010 in Pensacola and Choctawhatchee Bays, Florida, USA, to predict species distribution, field-collected and publically available data on oil distribution and concentration, and chronic toxicity data from laboratory assays applied to a matrix population model. The habitat suitability model established distribution of fish within Barataria Bay, Louisiana, USA, and the population model projected the dynamics of the species in the study area over a 5-yr period (October 2009-September 2014). Vital rates were modified according to estimated contaminant concentrations to simulate oil exposure effects. To evaluate the differences in levels of model complexity, simulations varied from temporally and spatially explicit, including seasonal variation and location-specific oiling, to simple interpretations of a risk quotient derived for the study area. The results of this study indicate that species distribution, as well as spatially and temporally variable contaminant concentrations, can provide a more ecologically relevant evaluation of species recovery from catastrophic environmental impacts but might not be cost-effective or efficient for rapid assessment needs.

A mechanism-based 3D-QSAR approach for classification and prediction of acetylcholinesterase inhibitory potency of organophosphate and carbamate analogs.

Organophosphate (OP) and carbamate esters can inhibit acetylcholinesterase (AChE) by binding covalently to a serine residue in the enzyme active site, and their inhibitory potency depends largely on affinity for the enzyme and the reactivity of the ester. Despite this understanding, there has been no mechanism-based in silico approach for classification and prediction of the inhibitory potency of ether OPs or carbamates. This prompted us to develop a three dimensional prediction framework for OPs, carbamates, and their analogs. Inhibitory structures of a compound that can form the covalent bond were identified through analysis of docked conformations of the compound and its metabolites. Inhibitory potencies of the selected structures were then predicted using a previously developed three dimensional quantitative structure-active relationship. This approach was validated with a large number of structurally diverse OP and carbamate compounds encompassing widely used insecticides and structural analogs including OP flame retardants and thio- and dithiocarbamate pesticides. The modeling revealed that: (1) in addition to classical OP metabolic activation, the toxicity of carbamate compounds can be dependent on biotransformation, (2) OP and carbamate analogs such as OP flame retardants and thiocarbamate herbicides can act as AChEI, (3) hydrogen bonds at the oxyanion hole is critical for AChE inhibition through the covalent bond, and (4) π-π interaction with Trp86 is necessary for strong inhibition of AChE. Our combined computation approach provided detailed understanding of the mechanism of action of OP and carbamate compounds and may be useful for screening a diversity of chemical structures for AChE inhibitory potency.

Bayesian Networks Improve Causal Environmental Assessments for Evidence-Based Policy.

Rule-based weight of evidence approaches to ecological risk assessment may not account for uncertainties and generally lack probabilistic integration of lines of evidence. Bayesian networks allow causal inferences to be made from evidence by including causal knowledge about the problem, using this knowledge with probabilistic calculus to combine multiple lines of evidence, and minimizing biases in predicting or diagnosing causal relationships. Too often, sources of uncertainty in conventional weight of evidence approaches are ignored that can be accounted for with Bayesian networks. Specifying and propagating uncertainties improve the ability of models to incorporate strength of the evidence in the risk management phase of an assessment. Probabilistic inference from a Bayesian network allows evaluation of changes in uncertainty for variables from the evidence. The network structure and probabilistic framework of a Bayesian approach provide advantages over qualitative approaches in weight of evidence for capturing the impacts of multiple sources of quantifiable uncertainty on predictions of ecological risk. Bayesian networks can facilitate the development of evidence-based policy under conditions of uncertainty by incorporating analytical inaccuracies or the implications of imperfect information, structuring and communicating causal issues through qualitative directed graph formulations, and quantitatively comparing the causal power of multiple stressors on valued ecological resources. These aspects are demonstrated through hypothetical problem scenarios that explore some major benefits of using Bayesian networks for reasoning and making inferences in evidence-based policy.

Acute Toxicity Prediction to Threatened and Endangered Species Using Interspecies Correlation Estimation (ICE) Models.

Evaluating contaminant sensitivity of threatened and endangered (listed) species and protectiveness of chemical regulations often depends on toxicity data for commonly tested surrogate species. The U.S. EPA's Internet application Web-ICE is a suite of Interspecies Correlation Estimation (ICE) models that can extrapolate species sensitivity to listed taxa using least-squares regressions of the sensitivity of a surrogate species and a predicted taxon (species, genus, or family). Web-ICE was expanded with new models that can predict toxicity to over 250 listed species. A case study was used to assess protectiveness of genus and family model estimates derived from either geometric mean or minimum taxa toxicity values for listed species. Models developed from the most sensitive value for each chemical were generally protective of the most sensitive species within predicted taxa, including listed species, and were more protective than geometric means models. ICE model estimates were compared to HC5 values derived from Species Sensitivity Distributions for the case study chemicals to assess protectiveness of the two approaches. ICE models provide robust toxicity predictions and can generate protective toxicity estimates for assessing contaminant risk to listed species.

Effects of Louisiana crude oil on the sheepshead minnow (Cyprinodon variegatus) during a life-cycle exposure to laboratory oiled sediment.

Determining the long-term effects of crude oil exposure is critical for ascertaining population-level ecological risks of spill events. A 19-week complete life-cycle experiment was conducted with the estuarine sheepshead minnow (Cyprinodon variegatus) exposed to reference (uncontaminated) sediment spiked with laboratory weathered South Louisiana crude (SLC) oil at five concentrations as well as one unspiked sediment control and one seawater (no sediment) control. Newly hatched larvae were exposed to the oiled sediments at measured concentrations of < 1 (sediment control), 50, 103, 193, 347, and 711 mg total polyaromatic hydrocarbons (tPAH)/kg dry sediment. Juveniles were exposed through the reproductively active adult phase at measured concentrations of <1 (sediment control), 52, 109, 199, 358, and 751 mg tPAH/kg sediment. Throughout the exposure, fish were assessed for growth, survival, and reproduction. Resulting F1 embryos were then collected, incubated, and hatched in clean water to determine if parental full life-cycle exposure to oiled sediment produced trans-generational effects. Larvae experienced significantly reduced standard length (5-13% reduction) and wet weight (13-35% reduction) at concentrations at and above 50 and 103 mg tPAH/kg sediment, respectively. At 92 and 132 days post hatch (dph), standard length was reduced (7-13% reduction) at 199 and 109 mg tPAH/kg dry sediment, respectively, and wet weight for both time periods was reduced at concentrations at and above 109 mg tPAH/kg dry sediment (21-38% reduction). A significant reduction (51-65%) in F0 fecundity occurred at the two highest test concentrations, but no difference was observed in F1 embryo survival. This study is the first to report the effects of chronic laboratory exposure to oiled sediment, and will assist the development of population models for evaluating risk to benthic spawning fish species exposed to oiled sediments. © 2015 Wiley Periodicals, Inc. Environ Toxicol 31: 1627-1639, 2016.

Determination of optimal timing of serial in-utero transfusions in red-cell alloimmunization.

OBJECTIVES: To assess the performance of middle cerebral artery peak systolic velocity (MCA-PSV) and of the expected daily decrease in fetal hemoglobin in determining the timing of serial in-utero transfusions (IUT) in red-cell alloimmunization. METHODS: This was a retrospective study of a continuous series of suspected anemic fetuses undergoing IUT between June 2003 and December 2012. Doppler measurement of MCA-PSV and pre- and post-transfusion hemoglobin levels were recorded at the time of the first, second and third IUT. Receiver-operating characteristics (ROC) curves and negative and positive predictive values of MCA-PSV in the prediction of severe fetal anemia were calculated. The daily decrease of fetal hemoglobin (Hb) between IUTs was calculated. Regression analysis was used to assess the correlation between pretransfusion fetal hemoglobin and MCA-PSV, and between observed and expected (by projection of daily decreases) pretransfusion fetal hemoglobin levels. RESULTS: One hundred and eleven fetuses required an IUT, of which 96 and 67 received a second and third IUT, respectively. The area under the ROC curve for MCA-PSV in the prediction of severe fetal anemia was not different for each rank of transfusion. The positive predictive value of MCA-PSV decreased from 75.3% at the first IUT, to 46.7% and 48.8% at the second and third IUTs, respectively, while the negative predictive value for a 1.5-MoM threshold remained high (88.9% at the second and 91.7% at the third IUT). The mean daily decrease in hemoglobin following each transfusion was 0.45, 0.35 and 0.32 g/dL, respectively. There was a persistent linear correlation between fetal hemoglobin and MCA-PSV and between observed and expected fetal hemoglobin levels. CONCLUSIONS: Both MCA-PSV and projection of daily decrease in hemoglobin are reliable means of diagnosing fetal anemia following previous IUTs. The high negative predictive value of MCA-PSV could allow subsequent IUTs to be postponed in selected cases.

Development and practical application of petroleum and dispersant interspecies correlation models for aquatic species.

Assessing the acute toxicity of oil has generally relied on existing toxicological data for a relatively few standard test species, which has limited the ability to estimate the impacts of spilled oil on aquatic communities. Interspecies correlation estimation (ICE) models were developed for petroleum and dispersant products to facilitate the prediction of toxicity values to a broader range of species and to better understand taxonomic differences in species sensitivity. ICE models are log linear regressions that can be used to estimate toxicity to a diversity of taxa based on the known toxicity value for a surrogate tested species. ICE models have only previously been developed for nonpetroleum chemicals. Petroleum and dispersant ICE models were statistically significant for 93 and 16 unique surrogate-predicted species pairs, respectively. These models had adjusted coefficient of determinations (adj-R(2)), square errors (MSE) and positive slope ranging from 0.29 to 0.99, 0.0002 to 0.311, and 0.187 to 2.665, respectively. Based on model cross-validation, predicted toxicity values for most ICE models (>90%) were within 5-fold of the measured values, with no influence of taxonomic relatedness on prediction accuracy. A comparison between hazard concentrations (HC) derived from empirical and ICE-based species sensitivity distributions (SSDs) showed that HC values were within the same order of magnitude of each other. These results show that ICE-based SSDs provide a statistically valid approach to estimating toxicity to a range of petroleum and dispersant products with applicability to oil spill assessment.

Developmental toxicity of Louisiana crude oil-spiked sediment to zebrafish.

Embryonic exposures to the components of petroleum, including polycyclic aromatic hydrocarbons (PAHs), cause a characteristic suite of developmental defects and cardiotoxicity in a variety of fish species. We exposed zebrafish embryos to reference sediment mixed with laboratory weathered South Louisiana crude oil and to sediment collected from an oiled site in Barataria Bay, Louisiana in December 2010. Laboratory oiled sediment exposures caused a reproducible set of developmental malformations in zebrafish embryos including yolk sac and pericardial edema, craniofacial and spinal defects, and tissue degeneration. Dose-response studies with spiked sediment showed that total polycyclic aromatic hydrocarbons (tPAH) concentrations of 27mg tPAH/kg (dry weight normalized to 1 percent organic carbon [1 percent OC]) caused a significant increase in defects, and concentrations above 78mg tPAH/kg 1 percent OC caused nearly complete embryo mortality. No toxicity was observed in Barataria sediment with 2mg tPAH/kg 1 percent OC. Laboratory aging of spiked sediment at 4°C resulted in a nearly 10-fold decrease in sensitivity over a 40-day period. This study demonstrates oiled sediment as an exposure pathway to fish with dose-dependent effects on embryogenesis that are consistent with PAH mechanisms of developmental toxicity. The results have implications for effects on estuarine fish from oiled coastal areas during the Deepwater Horizon spill.

Conserving large populations of lions - the argument for fences has holes.

Packer et al. reported that fenced lion populations attain densities closer to carrying capacity than unfenced populations. However, fenced populations are often maintained above carrying capacity, and most are small. Many more lions are conserved per dollar invested in unfenced ecosystems, which avoid the ecological and economic costs of fencing.

Prediction of aquatic toxicity mode of action using linear discriminant and random forest models.

The ability to determine the mode of action (MOA) for a diverse group of chemicals is a critical part of ecological risk assessment and chemical regulation. However, existing MOA assignment approaches in ecotoxicology have been limited to a relatively few MOAs, have high uncertainty, or rely on professional judgment. In this study, machine based learning algorithms (linear discriminant analysis and random forest) were used to develop models for assigning aquatic toxicity MOA. These methods were selected since they have been shown to be able to correlate diverse data sets and provide an indication of the most important descriptors. A data set of MOA assignments for 924 chemicals was developed using a combination of high confidence assignments, international consensus classifications, ASTER (ASessment Tools for the Evaluation of Risk) predictions, and weight of evidence professional judgment based an assessment of structure and literature information. The overall data set was randomly divided into a training set (75%) and a validation set (25%) and then used to develop linear discriminant analysis (LDA) and random forest (RF) MOA assignment models. The LDA and RF models had high internal concordance and specificity and were able to produce overall prediction accuracies ranging from 84.5 to 87.7% for the validation set. These results demonstrate that computational chemistry approaches can be used to determine the acute toxicity MOAs across a large range of structures and mechanisms.