| 1. |
- Liu, Tao, et al.
(author)
-
Modelling of partition constants : Linear solvation energy relationships or PLS regression?
- 2009
-
In: Journal of Chemometrics. - New York : Wiley. - 0886-9383 .- 1099-128X. ; 23:5, s. 254-262
-
Journal article (peer-reviewed)abstract
- Estimation methods for partition constants are needed in many fields of engineering and science. The partitioning between phases is determined by the free energy of the transfer and all estimation methods must therefore describe the same entity. Linear solvation energy relationships (LSERs) try to split the contributions to van der Waals and polar interactions into directly interpretable solute descriptors, while projection-based regression methods can accomplish a similar dimensionality reduction from a set of theoretical descriptors. Here, we use the partitioning between octanol and water (Kow) and water solubility (Sw) to investigate similarities and differences between LSER and partial least squares regression (PLSR) models. The similarities in model structure are described, and shown to transform into a comparable prediction performance. We also demonstrate the opportunity to accomplish an analogous chemical interpretation of a PLSR model - either directly or through a linear transformation of the PLS factors - as with an LSER model. Much of the alleged difference between the mechanistic or semi-empirical LSER and the statistical PLSR models will then disappear. The choice of a modelling approach should therefore primarily be driven by the availability of data and predictive performance.
|
|
| 2. |
|
|
| 3. |
- Öberg, Tomas, 1956-
(author)
-
Probabilistisk riskbedömning fas 1. : Sannolikhetsbaserad uppskattning av miljö- och hälsorisker i förorenade markområden – en litteraturöversikt.
- 2006
-
Reports (other academic/artistic)abstract
- Förorenad mark medför risker för hälsa och miljö. Litteraturöversikten redovisar hur nuvarande metodik för kvantitativa riskbedömningar kan kompletteras med en sannolikhetsbaserad – probabilistisk – ansats. Den probabilistiska metoden innebär att variabilitet (naturlig variation) och osäkerhet (okunskap) kan karakteriseras och därmed erhålls både ett bättre beslutsunderlag och kunskap om hur bedömningen ytterligare kan förfinas. Den probabilistiska metodiken har fått stor användning i USA och nu börjar den även tillämpas i flera europeiska länder. Probabilistiska riskbedömningar baseras i allmänhet på simuleringar av utfall från ett stort antal möjliga val av värden för ingångsvariabler och modellparametrar. Beräkningarna kan numer utföras med en vanlig persondator, men kräver en grundläggande kunskap om både teknikens möjligheter och begränsningar. Förorenad mark är vid sidan av kärnkraftsindustrin den viktigaste miljötillämpningen av probabilistisk riskbedömning och ett stort antal studier har publicerats för specifika objekt i Nordamerika, Europa och Asien. De avser exempelvis förorening med bly, arsenik, krom, uran, PCB, PAH, hexaklorbensen, pentaklorfenol, dioxiner och klorerade lösningsmedel. Dessa probabilistiska riskbedömningar täcker in olika exponeringssituationer inom vitt skilda verksamheter, däribland tidigare metallurgisk industri (smältverk och gruvor), tillverkningsindustri, gasverkstomter, träimpregnering, infrastruktur och deponier. En övergång till probabilistisk riskbedömning ställer krav på kvalitetssäkring, både avseende arbetsgången och redovisningsrutinerna. Det amerikanska naturvårdsverket (U. S. EPA) har gett ut relativt detaljerade anvisningar som i stort överensstämmer med vad som idag utgör vetenskapligt konsensus. Liknande behov av riktlinjer finns även i Europa. Behovet av att karakterisera variabilitet, osäkerhet och känslighet i riskbedömningsmodeller är inte annorlunda i Sverige än i Nordamerika. Dessutom behöver säkerhetsmarginalernas storlek klart kunna anges. Probabilistisk metodik kan enkelt integreras med nuvarande svenska riskbedömningsmodeller och rapporten redovisar ett beräkningsexempel för benso[a]pyren. Probabilistisk riskbedömning har ofta använts för att etablera platsspecifika riktvärden och det är här som en framtida användning i Sverige främst kan förutses. Rapporten pekar på behovet av ramar för att underlätta tolkning och jämförbarhet och rekommenderar att ett vägledningsdokument utarbetas. Likaså krävs utbildningsinsatser. Kurser i probabilistiska metoder finns i det ordinarie utbildningsprogrammet vid ett par högskolor, men ett behov finns även av kurser för fortbildning av redan yrkesverksamma, både som distansutbildning och kortare problembaserade kurser.
|
|
| 4. |
- Cederqvist, Lars, et al.
(author)
-
Improved process stability during friction stir welding of 5 cm thick copper canisters through shoulder geometry and parameter studies
- 2009
-
In: Science and technology of welding and joining. - London : Institute of Materials. - 1362-1718 .- 1743-2936. ; 14:2, s. 178-184
-
Journal article (peer-reviewed)abstract
- The spent nuclear fuel from Swedish power plants will be placed in copper canisters that are sealed with friction stir welding and the stability and robustness of this process is now being optimised in three steps: first, the shoulder geometry was identified that produced the most stable weld cycle, then the welding parameters were optimised for that geometry with regards to stability, and finally, the chosen geometry and welding parameters were verified and evaluated during multiple weld cycles. The shoulder study showed that stable welds could be produced repeatedly with a convex scroll geometry which proved more stable than various concave and flat scroll geometries. In the subsequent parameter study, not only were the most stable values for the welding parameters derived, but a clear relationship was shown between power input and tool temperature. This relationship can be used to more accurately control the process within the parameter windows, not only for this application but for other applications where the welding temperature needs to be kept within a specified range. Similarly, the potential of the convex scroll shoulder geometry for use in applications with other metals and thicknesses is evident.
|
|
| 5. |
- Sander, Per, et al.
(author)
-
Comparing deterministic and probabilistic risk assessments : A case study at a closed steel mill in southern Sweden
- 2006
-
In: Journal of Soils and Sediments. - Landsberg : Ecomed Pub.. - 1439-0108 .- 1614-7480. ; 6:1, s. 55-61
-
Journal article (peer-reviewed)abstract
- Background, Aims and Scope. Contaminated land is a high priority environmental problem in most of Europe and North-America. Sweden is no exception and generic guideline values have been developed for the initial assessment, but site-specific assessments are also needed. The generic guideline values are not applicable when the exposure conditions are different from the typical Swedish conditions or when the site contains a particularly sensitive ecosystem. The Swedish guideline values have, like in many other countries, been set by using deterministic point estimates for all variables and constants in the used multimedia model. The same approach is common also for site-specific assessments, and a limitation is that it fails to quantify variability and uncertainty. Probabilistic risk assessment provided a method to deal with this problem. Variability and uncertainty in the input parameters (variables or constants) are described by probability distributions, and likewise the output (risk or exposure) is presented as a probability distribution. A substantial number of probabilistic risk assessments for contaminated land at sites in North America, Europe and Asia have been published. However, an extensive review of the literature did not identify any study where probabilistic risk assessment was applied to a site contaminated by an iron or steel industry. Here we will describe such a case, where we have compared a deterministic point estimate with a probabilistic risk assessment for six elements and benzo[a]pyrene. Methods. The site had different metallurgical plants in operation for more than 100 years. Most parts of the steel mill were closed by the mid 1980s, and today the site is used by small-sized enterprises. The soil is contaminated with metals from the previous industrial operations. The present owner plans to develop the site and has therefore initiated extensive investigations of soil contamination. Sixty-two soil samples collected between 1997 and 2000 provided a good coverage of the whole site, and were analyzed for the content of different elements and polycyclic aromatic hydrocarbons (PAH). The exposure assessments were focused on six elements with high concentrations compared to the generic guideline values; arsenic (As), lead (Pb), cadmium (Cd), chromium (Cr), copper (Cu) and zinc (Zn). In addition, benzo[a]pyrene was included due to the high toxicity and comparatively high concentrations. Variability and uncertainty were characterized in a Monte Carlo simulation of exposures (10 000 iterations), and the exposures were evaluated with two land use scenarios; less sensitive use and sensitive use. Results and Discussion. The deterministic point estimates and the probabilistic estimates of the 95th percentile are in approximately the same ranges in the scenario of less sensitive land use. It is only the exposure for arsenic that is slightly above the toxicological reference value (TRV) in the deterministic assessment. In the probabilistic assessment, the exposure for all elements is below the TRV. The results for sensitive land use are applicable to a scenario where the site is developed for general housing. The deterministic point estimates and the probabilistic estimates of the 95th percentile are also here in approximately the same ranges, but the exposure exceeds the TRV for arsenic, cadmium and lead. Drinking water, vegetables grown on site and soil ingestion are the major exposure pathways for this scenario. In this assessment, the estimated intake distributions are applicable to a randomly selected individual. The probability distributions used here to characterize the different soil parameters are typically representing both variability and uncertainty, and the same is true the majority of the exposure variables. We therefore decided not to attempt to separate variability and uncertainty at this stage, but with additional data from a more in-depth site investigation it might be possible to achieve this. Conclusions and Outlook. To the best of our knowledge, this study is the first report on a probabilistic risk assessment on a former iron and steel works site. The materials handled by this industry were less toxic than for many other metallurgical operations, but contaminants may still severely limit the options for future land use. This case study shows that probabilistic exposure estimates for a set of soil contaminants can be quite similar to deterministic point estimates. The main difference is instead to be found in the additional information obtained with the probabilistic assessment. The sensitivity analyses show pathways and input variables that contribute most to variations in the total intake of each contaminant, e.g. dermal contact and ingestion of soil, vegetables and drinking water. This information can be used both in the planning of future land use and for active measures to reduce current exposure. The probabilistic assessment also provides information on the magnitude of exposure and the margin of safety. This information may facilitate risk communication between decision-makers and stakeholders. The presentation of results from probabilistic risk assessments is only briefly discussed in the literature and here we see a need for research and opportunities for enhancement. The choice of data analytical tools may then be of importance, since more complex multimedia models are rather difficult to decipher when implemented within traditional spreadsheet software. Some of the research needs are identified here and in a previous review article in this journal.
|
|
| 6. |
- Sander, Per, et al.
(author)
-
Uncertain numbers and uncertainty in the selection of input distributions--consequences for a probabilistic risk assessment of contaminated land.
- 2006
-
In: Risk Analysis. - New York : Plenum Press. - 0272-4332 .- 1539-6924. ; 26:5, s. 1363-1375
-
Journal article (peer-reviewed)abstract
- Risks from exposure to contaminated land are often assessed with the aid of mathematical models. The current probabilistic approach is a considerable improvement on previous deterministic risk assessment practices, in that it attempts to characterize uncertainty and variability. However, some inputs continue to be assigned as precise numbers, while others are characterized as precise probability distributions. Such precision is hard to justify, and we show in this article how rounding errors and distribution assumptions can affect an exposure assessment. The outcome of traditional deterministic point estimates and Monte Carlo simulations were compared to probability bounds analyses. Assigning all scalars as imprecise numbers (intervals prescribed by significant digits) added uncertainty to the deterministic point estimate of about one order of magnitude. Similarly, representing probability distributions as probability boxes added several orders of magnitude to the uncertainty of the probabilistic estimate. This indicates that the size of the uncertainty in such assessments is actually much greater than currently reported. The article suggests that full disclosure of the uncertainty may facilitate decision making in opening up a negotiation window. In the risk analysis process, it is also an ethical obligation to clarify the boundary between the scientific and social domains.
|
|
| 7. |
- Tetko, Igor V, et al.
(author)
-
Critical assessment of QSAR models of environmental toxicity against Tetrahymena pyriformis : focusing on applicability domain and overfitting by variable selection.
- 2008
-
In: Journal of chemical information and modeling. - : American Chemical Society. - 1549-9596 .- 1549-960X. ; 48:9, s. 1733-46
-
Journal article (peer-reviewed)abstract
- The estimation of the accuracy of predictions is a critical problem in QSAR modeling. The "distance to model" can be defined as a metric that defines the similarity between the training set molecules and the test set compound for the given property in the context of a specific model. It could be expressed in many different ways, e.g., using Tanimoto coefficient, leverage, correlation in space of models, etc. In this paper we have used mixtures of Gaussian distributions as well as statistical tests to evaluate six types of distances to models with respect to their ability to discriminate compounds with small and large prediction errors. The analysis was performed for twelve QSAR models of aqueous toxicity against T. pyriformis obtained with different machine-learning methods and various types of descriptors. The distances to model based on standard deviation of predicted toxicity calculated from the ensemble of models afforded the best results. This distance also successfully discriminated molecules with low and large prediction errors for a mechanism-based model developed using log P and the Maximum Acceptor Superdelocalizability descriptors. Thus, the distance to model metric could also be used to augment mechanistic QSAR models by estimating their prediction errors. Moreover, the accuracy of prediction is mainly determined by the training set data distribution in the chemistry and activity spaces but not by QSAR approaches used to develop the models. We have shown that incorrect validation of a model may result in the wrong estimation of its performance and suggested how this problem could be circumvented. The toxicity of 3182 and 48774 molecules from the EPA High Production Volume (HPV) Challenge Program and EINECS (European chemical Substances Information System), respectively, was predicted, and the accuracy of prediction was estimated. The developed models are available online at http://www.qspr.org site.
|
|
| 8. |
- Zhu, Hao, et al.
(author)
-
Combinatorial QSAR modeling of chemical toxicants tested against Tetrahymena pyriformis
- 2008
-
In: Journal of chemical information and modeling. - : American Chemical Society. - 1549-9596 .- 1549-960X. ; 48:4, s. 766-784
-
Journal article (peer-reviewed)abstract
- Selecting most rigorous quantitative structure-activity relationship (QSAR) approaches is of great importance in the development of robust and predictive models of chemical toxicity. To address this issue in a systematic way, we have formed an international virtual collaboratory consisting of six independent groups with shared interests in computational chemical toxicology. We have compiled an aqueous toxicity data set containing 983 unique compounds tested in the same laboratory over a decade against Tetrahymena pyriformis. A modeling set including 644 compounds was selected randomly from the original set and distributed to all groups that used their own QSAR tools for model development. The remaining 339 compounds in the original set (external set I) as well as 110 additional compounds (external set II) published recently by the same laboratory (after this computational study was already in progress) were used as two independent validation sets to assess the external predictive power of individual models. In total, our virtual collaboratory has developed 15 different types of QSAR models of aquatic toxicity for the training set. The internal prediction accuracy for the modeling set ranged from 0.76 to 0.93 as measured by the leave-one-out cross-validation correlation coefficient ( Q abs2). The prediction accuracy for the external validation sets I and II ranged from 0.71 to 0.85 (linear regression coefficient R absI2) and from 0.38 to 0.83 (linear regression coefficient R absII2), respectively. The use of an applicability domain threshold implemented in most models generally improved the external prediction accuracy but at the same time led to a decrease in chemical space coverage. Finally, several consensus models were developed by averaging the predicted aquatic toxicity for every compound using all 15 models, with or without taking into account their respective applicability domains. We find that consensus models afford higher prediction accuracy for the external validation data sets with the highest space coverage as compared to individual constituent models. Our studies prove the power of a collaborative and consensual approach to QSAR model development. The best validated models of aquatic toxicity developed by our collaboratory (both individual and consensus) can be used as reliable computational predictors of aquatic toxicity and are available from any of the participating laboratories.
|
|
| 9. |
- Öberg, Tomas, 1956-
(author)
-
A general structure-property relationship to predict the enthalpy of vaporisation at ambient temperatures.
- 2007
-
In: SAR and QSAR in environmental research (Print). - Reading, Berkshire, Great Britain : Gordon and Breach Science Publishers. - 1062-936X .- 1029-046X. ; 18:1-2, s. 127-139
-
Journal article (peer-reviewed)abstract
- The vapour pressure is the most important property of an anthropogenic organic compound in determining its partitioning between the atmosphere and the other environmental media. The enthalpy of vaporisation quantifies the temperature dependence of the vapour pressure and its value around 298 K is needed for environmental modelling. The enthalpy of vaporisation can be determined by different experimental methods, but estimation methods are needed to extend the current database and several approaches are available from the literature. However, these methods have limitations, such as a need for other experimental results as input data, a limited applicability domain, a lack of domain definition, and a lack of predictive validation. Here we have attempted to develop a quantitative structure-property relationship (QSPR) that has general applicability and is thoroughly validated. Enthalpies of vaporisation at 298 K were collected from the literature for 1835 pure compounds. The three-dimensional (3D) structures were optimised and each compound was described by a set of computationally derived descriptors. The compounds were randomly assigned into a calibration set and a prediction set. Partial least squares regression (PLSR) was used to estimate a low-dimensional QSPR model with 12 latent variables. The predictive performance of this model, within the domain of application, was estimated at n=560, q2Ext=0.968 and s=0.028 (log transformed values). The QSPR model was subsequently applied to a database of 100,000+ structures, after a similar 3D optimisation and descriptor generation. Reliable predictions can be reported for compounds within the previously defined applicability domain.
|
|
| 10. |
- Öberg, Tomas, 1956-
(author)
-
A QSAR for the hydroxyl radical reaction rate constant : Validation, domain of application, and prediction
- 2005
-
In: Atmospheric Environment. - : Elsevier BV. - 1352-2310 .- 1873-2844. ; 39:12, s. 2189-2200
-
Journal article (peer-reviewed)abstract
- A large number of anthropogenic organic chemicals are emitted into the troposphere. Reactions with the hydroxyl radical are a dominant removal pathway for most organic compounds, but experimentally determined gas-phase reaction rate constants are only available for about 750 compounds. The lack of experimental data increases the importance of applying structure-activity relationships (QSAR) to evaluate and predict reactivities. It is generally acknowledged that these empirical relationships are valid only within the same domain for which they were developed. However, model validation is sometimes neglected and the application domain is not always well defined. The purpose of this paper is to outline how validation and domain definition can facilitate the modeling and prediction of the hydroxyl radical reaction rates for a large database. A substantial number of theoretical descriptors (867) were generated from 2D molecular structures for compounds present in the Syracuse Research Corporationﾒs PhysProp Database. A QSAR model was developed for the hydroxyl radical reaction rate constant using a projection-based regression technique, PLSR (partial least squares regression). The PLSR model was subsequently validated with an external test set. The main factors of variation could be attributed to two reaction pathways, hydrogen atom abstraction and addition to double bonds or aromatic systems. A set of 17 293 compounds, drawn from the PhysProp Database, was projected onto the PLSR model and 74% were inside the applicability domain. The predicted hydroxyl reaction rates for 25% of these compounds were slow or negligible, with atmospheric half-lives in the range from days to years. Finally, the list of persistent organic compounds was matched against the OECD list of High Production Volume Chemicals (HPVC). Together with the experimental data, nearly three hundred compounds were identified as both persistent and in high volume production.
|
|
