A Risk Assessment Perspective of Current Practice in Characterizing Uncertainties in QSAR Regression Predictions
(2011) In Molecular Informatics 30(6-7). p.551-564- Abstract
- The European REACH legislation accepts the use of non-testing methods, such as QSARs, to inform chemical risk assessment. In this paper, we aim to initiate a discussion on the characterization of predictive uncertainty from QSAR regressions. For the purpose of decision making, we discuss applications from the perspective of applying QSARs to support probabilistic risk assessment. Predictive uncertainty is characterized by a wide variety of methods, ranging from pure expert judgement based on variability in experimental data, through data-driven statistical inference, to the use of probabilistic QSAR models. Model uncertainty is dealt with by assessing confidence in predictions and by building consensus models. The characterization of... (More)
- The European REACH legislation accepts the use of non-testing methods, such as QSARs, to inform chemical risk assessment. In this paper, we aim to initiate a discussion on the characterization of predictive uncertainty from QSAR regressions. For the purpose of decision making, we discuss applications from the perspective of applying QSARs to support probabilistic risk assessment. Predictive uncertainty is characterized by a wide variety of methods, ranging from pure expert judgement based on variability in experimental data, through data-driven statistical inference, to the use of probabilistic QSAR models. Model uncertainty is dealt with by assessing confidence in predictions and by building consensus models. The characterization of predictive uncertainty would benefit from a probabilistic formulation of QSAR models (e. g. generalized linear models, conditional density estimators or Bayesian models). This would allow predictive uncertainty to be quantified as probability distributions, such as Bayesian predictive posteriors, and likelihood-based methods to address model uncertainty. QSAR regression models with point estimates as output may be turned into a probabilistic framework without any loss of validity from a chemical point of view. A QSAR model for use in probabilistic risk assessment needs to be validated for its ability to make reliable predictions and to quantify associated uncertainty. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/3800124
- author
- Sahlin, Ullrika LU ; Filipsson, Monika and Oberg, Tomas
- publishing date
- 2011
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- Decision making, Predictive uncertainty, Probabilistic risk assessment, REACH, Regression
- in
- Molecular Informatics
- volume
- 30
- issue
- 6-7
- pages
- 551 - 564
- publisher
- John Wiley & Sons Inc.
- external identifiers
-
- wos:000293847700006
- scopus:79960573919
- pmid:27467156
- ISSN
- 1868-1751
- DOI
- 10.1002/minf.201000177
- language
- English
- LU publication?
- no
- id
- 772947c2-ec1b-41da-8aed-6ed0b7b9053b (old id 3800124)
- date added to LUP
- 2016-04-01 09:54:51
- date last changed
- 2022-01-25 17:52:54
@article{772947c2-ec1b-41da-8aed-6ed0b7b9053b,
abstract = {{The European REACH legislation accepts the use of non-testing methods, such as QSARs, to inform chemical risk assessment. In this paper, we aim to initiate a discussion on the characterization of predictive uncertainty from QSAR regressions. For the purpose of decision making, we discuss applications from the perspective of applying QSARs to support probabilistic risk assessment. Predictive uncertainty is characterized by a wide variety of methods, ranging from pure expert judgement based on variability in experimental data, through data-driven statistical inference, to the use of probabilistic QSAR models. Model uncertainty is dealt with by assessing confidence in predictions and by building consensus models. The characterization of predictive uncertainty would benefit from a probabilistic formulation of QSAR models (e. g. generalized linear models, conditional density estimators or Bayesian models). This would allow predictive uncertainty to be quantified as probability distributions, such as Bayesian predictive posteriors, and likelihood-based methods to address model uncertainty. QSAR regression models with point estimates as output may be turned into a probabilistic framework without any loss of validity from a chemical point of view. A QSAR model for use in probabilistic risk assessment needs to be validated for its ability to make reliable predictions and to quantify associated uncertainty.}},
author = {{Sahlin, Ullrika and Filipsson, Monika and Oberg, Tomas}},
issn = {{1868-1751}},
keywords = {{Decision making; Predictive uncertainty; Probabilistic risk assessment; REACH; Regression}},
language = {{eng}},
number = {{6-7}},
pages = {{551--564}},
publisher = {{John Wiley & Sons Inc.}},
series = {{Molecular Informatics}},
title = {{A Risk Assessment Perspective of Current Practice in Characterizing Uncertainties in QSAR Regression Predictions}},
url = {{http://dx.doi.org/10.1002/minf.201000177}},
doi = {{10.1002/minf.201000177}},
volume = {{30}},
year = {{2011}},
}