The importance of outlier detection and training set selection for reliable environmental QSAR predictions

Furusjö, Erik; Svenson, Anders; Rahmberg, Magnus; Andersson, Magnus

doi:10.1016/j.chemosphere.2005.07.002

Cited by 79 publications

(48 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…A partial least squares regression model was developed for several environmental endpoints, including acute toxicity to fish, daphnid and algae [88]. The paper illustrates different methods for diagnostics that can be used to decide whether or not a substance is within the model domain.…”

Section: Examples Of Statistically Derived Qsarsmentioning

confidence: 99%

Review of (Quantitative) Structure–Activity Relationships for Acute Aquatic Toxicity

2008

View full text Add to dashboard Cite

This paper reviews different approaches described in the literature for estimating the aquatic toxicity of chemical substances. It is based on an extended review performed by the European Chemicals Bureau of the European Commissions Joint Research Centre in support of the development of technical guidance for the implementation of the REACH legislation, and is one of a series of minireviews in this journal. The paper is organised by approach for (Q)SAR development and includes a review of: (i) (Q)SARs for acute aquatic toxicity by chemical class, (ii) (Q)SARs for acute aquatic toxicity by mode of action, (iii) review of statistically derived (Q)SARs, (iv) structural alerts for excess aquatic toxicity and (v) expert systems that combine structural rules and multiple (Q)SAR models to predict aquatic toxicological endpoints. Directions and recommendations for further research are also provided.

show abstract

Section: Examples Of Statistically Derived Qsarsmentioning

confidence: 99%

Review of (Quantitative) Structure–Activity Relationships for Acute Aquatic Toxicity

2008

View full text Add to dashboard Cite

show abstract

“…20,21 Therefore, some outliers detection methods [22][23][24][25] such as student residence and leverage methods were used to remove outliers. According to this method, no samples in both the 1 and 4 mm pathlength model were considered anomalous in the full spectral regions evaluated, thereby remaining the dataset with 90 samples.…”

Section: Outlier Detectionmentioning

confidence: 99%

The relevance study of effective information between near infrared spectroscopy and chondroitin sulfate in ethanol precipitation process

Ding

et al. 2014

J. Innov. Opt. Health Sci.

View full text Add to dashboard Cite

Near infrared spectroscopy (NIRS) is based on molecular overtone and combination vibrations. It is di±cult to assign speci¯c features under complicated system. So it is necessary to¯nd the relevance between NIRS and target compound. For this purpose, the chondroitin sulfate (CS) ethanol precipitation process was selected as the research model, and 90 samples of 5 di®erent batches were collected and the content of CS was determined by modi¯ed carbazole method. The relevance between NIRS and CS was studied throughout optical pathlength, pretreatment methods and variables selection methods. In conclusion, the¯rst derivative with Savitzky-Golay (SG) smoothing was selected as the best pretreatment, and the best spectral region was selected using interval partial least squares (iPLS) method under 1 mm optical cell. A multivariate calibration model was established using PLS algorithm for determining the content of CS, and the root mean square error of prediction (RMSEP) is 3.934 gÁL À1 . This method will have great potential in process analytical technology in the future.

show abstract

“…In multivariate statistics, it is common to define three types of outliers. 23) 1. X/Y relation outliers are substances for which the relationship between the descriptors (X variables) and the dependent variables (Y variables) is not the same as in the (rest of the) training data.…”

Section: Methodsmentioning

confidence: 99%

A QSAR Approach for the Prediction of Stability of Benzoglycolamide Ester Prodrugs

Narasimhan

Ansari²,

Singh

et al. 2006

CHEMICAL & PHARMACEUTICAL BULLETIN

View full text Add to dashboard Cite

Quantitative structure-activity relationship (QSAR) represents an attempt to relate structural descriptors of compounds with their physicochemical properties and biological activities. It is widely used for the prediction of physicochemical properties in chemical, environmental, and pharmaceutical areas.1,2) The main steps involved in this method include the following: data collection, molecular descriptor selection and procurement, correlation model development and finally model evaluation. At present, many types of molecular descriptors have been proposed to describe the structural features of the molecules. [3][4][5] The main focus of these studies is to search for regularity in the manner in which molecular properties change as molecular structure changes. 6,7) One such property is the hydrolysis of drug molecules. Double ester prodrugs of penicillins are more susceptible to enzymatic hydrolysis than their simple alkyl esters but have the problem of poor aqueous solubility. This problem of carboxylic acid prodrugs was overcome by the synthesis of benzoglycolamide esters.8) The literature reveals many reports on QSRR (quantitative structure reactivity relationship).9-15) Chung et al. 16) has shown structure reactivity relationships in chemical hydrolysis of prodrugs of nicotinic acid. Similar studies on QSRR were found in literature. [17][18][19] In such studies the reactivity, in terms of rate constants, is determined and then correlated with structural parameters by use of suitable statistical techniques. The aim of this study was to establish quantitative relationship(s) between structure of benzogylcolamide esters 8) and their basic hydrolytic rate constants. Bundgaard et al. 8) synthesized benzoglycolamide esters and studied their physicochemical properties and stability, which were based on qualitative observations, which can be fortuitous also. But the correlation of benzoglycolamide ester structure with hydrolytic rate constant has not yet been investigated.In view of above and in continuation of our studies on correlation of molecular properties with activity, [20][21][22] we decided to study the usefulness of QSAR in the prediction of the basic hydrolytic rate constant of benzoglycolamide esters synthesized by Bundgaard et al.8) The multiple linear regression (MLR) model has been developed as a calibration model for predicting the Ϫlog kc of compounds belonging to test set and investigating their linear characteristics. The present results are important from a point of accurate prediction of Ϫlog kc. Our model can predict the basic hydrolytic rate constant of variety of a benzoglycolamide esters and it reduces the need for different time consuming experiments. MethodologyThe main aim of the present work was development of a quantitative model to predict the basic hydrolytic rate constant of different benzoglycolamide esters. In the present work, the hydrolytic rate constants of 44 benzoglycolamide esters synthesized 8) were subjected to MLR with their physicochemical properties. The best MLR mod...

show abstract

The importance of outlier detection and training set selection for reliable environmental QSAR predictions

Cited by 79 publications

References 27 publications

Review of (Quantitative) Structure–Activity Relationships for Acute Aquatic Toxicity

Review of (Quantitative) Structure–Activity Relationships for Acute Aquatic Toxicity

The relevance study of effective information between near infrared spectroscopy and chondroitin sulfate in ethanol precipitation process

A QSAR Approach for the Prediction of Stability of Benzoglycolamide Ester Prodrugs

Contact Info

Product

Resources

About