Methods for assigning confidence to toxicity data with multiple values — Identifying experimental outliers

Steinmetz, Fabian P.; Enoch, Steven J.; Madden, Judith C.; Nelms, Mark; Rodríguez-Sánchez, Neus; Rowe, Phil; Wen, Yuezhong; Cronin, Mark T.D.

doi:10.1016/j.scitotenv.2014.02.115

Cited by 15 publications

(29 citation statements)

References 59 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…toxicity is increasing linearly with lipophilicity. 5,[22][23][24]43 Consideration of the QSARs developed in this study shows an improvement in the models when utilising CS-weighted regression. The improvement is both the statistical fit but also the slope for log K OW which approaches one when employing CS-weighting, i.e.…”

Section: Evaluation Of the Predictivity Of The Qsars/qsprsmentioning

confidence: 85%

“…The negative logarithm of the effective concentration causing 50% light reduction (EC 50 ) is expressed as the pT. 20 Extending the original compilation of Kaiser and Palabrica 20 , Steinmetz et al 5 collected a large meta-dataset with 1813 different values for Microtox toxicity. In order to create meaningful QSAR models in aquatic toxicology, there is an application of the well-established relationship between acute toxicity and hydrophobicity for compounds acting by the non-polar narcosis mechanism of action.…”

Section: Aquatic Toxicologymentioning

confidence: 99%

“…Any indication of the quality of data would be very helpful for purposes such as risk assessment, but more crucially for modelling including QSARs and read-across prediction. 3,5 Therefore, the aim of this study was to investigate how using approaches for statistical data quality, i.e. CS, improve the development of QSAR/QSPR models.…”

Section: Aquatic Toxicologymentioning

confidence: 99%

“…43,44 This study has expanded the number of chemicals 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 21 with data within non-polar narcosis domain for A. fischeri, hence expanding the chemical space and extended a previous study. 5 It is of no surprise that the non-polar narcosis data for A. fischeri (Microtox) are significantly correlated with log K OW , even if some historical data are obviously of quite poor quality. 5 The QSAR (Eq.…”

Section: Evaluation Of the Predictivity Of The Qsars/qsprsmentioning

confidence: 99%

See 3 more Smart Citations

Data Quality in the Human and Environmental Health Sciences: Using Statistical Confidence Scoring to Improve QSAR/QSPR Modeling

Steinmetz

Madden

Cronin

2015

J. Chem. Inf. Model.

Self Cite

View full text Add to dashboard Cite

ABSTRACT:A greater number of toxicity data are becoming publicly available allowing for in silico modelling. However, questions often arise as how to incorporate data quality and how to deal with contradicting data if more than a single datum point is available for the same compound. In this study, two well-known and studied QSAR/QSPR models for skin permeability and aquatic toxicology have been investigated in the context of statistical data quality. In particular, the potential benefits of the incorporation of the statistical Confidence Scoring (CS) approach within modelling and validation. As a result, robust QSAR/QSPR models for the skin permeability coefficient and the toxicity of non-polar narcotics to Aliivibrio fischeri assay were created. CSweighted linear regression for training and CS-weighted root mean square error (RMSE) for validation were statistically superior compared to standard linear regression and standard RMSE.Strategies are proposed as to how to interpret data with high and low CS, as well as how to deal with large datasets containing multiple entries.

show abstract

Section: Evaluation Of the Predictivity Of The Qsars/qsprsmentioning

confidence: 85%

Section: Aquatic Toxicologymentioning

confidence: 99%

Section: Aquatic Toxicologymentioning

confidence: 99%

Section: Evaluation Of the Predictivity Of The Qsars/qsprsmentioning

confidence: 99%

See 2 more Smart Citations

Data Quality in the Human and Environmental Health Sciences: Using Statistical Confidence Scoring to Improve QSAR/QSPR Modeling

Steinmetz

Madden

Cronin

2015

J. Chem. Inf. Model.

Self Cite

View full text Add to dashboard Cite

show abstract

“…The higher the experimental error within a data set, then the subsequent QSA(P)R model will be poorer based on the RMSE in prediction for a representative external test set [86,87]. For a QSA(P)R model to have a RMSE in prediction for a representative external test set of B0.30 then the experimental error associated with a data set needs to be B0.20 [86].…”

Section: [P] + [D]mentioning

confidence: 99%

Interpreting physicochemical experimental data sets

Colclough

Wenlock

2015

J Comput Aided Mol Des

View full text Add to dashboard Cite

With the wealth of experimental physicochemical data available to chemoinformaticians from the literature, commercial, and company databases an increasing challenge is the interpretation of such datasets. Subtle differences in experimental methodology used to generate these datasets can give rise to variations in physicochemical property values. Such methodology nuances will be apparent to an expert experimentalist but not necessarily to the data analyst and modeller. This paper describes the differences between common methodologies for measuring the four most important physicochemical properties namely aqueous solubility, octan-1-ol/water distribution coefficient, pK(a) and plasma protein binding highlighting key factors that can lead to systematic differences. Insight is given into how to identify datasets suitable for combining.

show abstract

In silico toxicology: From structure–activity relationships towards deep learning and adverse outcome pathways

Hemmerich

Ecker

2020

WIREs Comput Mol Sci

109

View full text Add to dashboard Cite

In silico toxicology is an emerging field. It gains increasing importance as research is aiming to decrease the use of animal experiments as suggested in the 3R principles by Russell and Burch. In silico toxicology is a means to identify hazards of compounds before synthesis, and thus in very early stages of drug development. For chemical industries, as well as regulatory agencies it can aid in gap‐filling and guide risk minimization strategies. Techniques such as structural alerts, read‐across, quantitative structure–activity relationship, machine learning, and deep learning allow to use in silico toxicology in many cases, some even when data is scarce. Especially the concept of adverse outcome pathways puts all techniques into a broader context and can elucidate predictions by mechanistic insights. This article is categorized under: Structure and Mechanism > Computational Biochemistry and Biophysics Data Science > Chemoinformatics

show abstract

Methods for assigning confidence to toxicity data with multiple values — Identifying experimental outliers

Cited by 15 publications

References 59 publications

Data Quality in the Human and Environmental Health Sciences: Using Statistical Confidence Scoring to Improve QSAR/QSPR Modeling

Data Quality in the Human and Environmental Health Sciences: Using Statistical Confidence Scoring to Improve QSAR/QSPR Modeling

Interpreting physicochemical experimental data sets

In silico toxicology: From structure–activity relationships towards deep learning and adverse outcome pathways

Contact Info

Product

Resources

About