Principles of QSAR models validation: internal and external

Wiley - Tập 26 Số 5 - Trang 694-701 - 2007
Paola Gramatica1
1QSAR Research Unit in Environmental Chemistry and Ecotoxicology, Department of Structural and Functional Biology, University of Insubria, Via Dunant 3, 21100 Varese, Italy

Tóm tắt

AbstractThe recent REACH Policy of the European Union has led to scientists and regulators to focus their attention on establishing general validation principles for QSAR models in the context of chemical regulation (previously known as the Setubal, nowadays, the OECD principles). This paper gives a brief analysis of some principles: unambiguous algorithm, Applicability Domain (AD), and statistical validation. Some concerns related to QSAR algorithm reproducibility and an example of a fast check of the applicability domain for MLR models are presented. Common myths and misconceptions related to popular techniques for verifying internal predictivity, particularly for MLR models (for instance cross‐validation, bootstrap), are commented on and compared with commonly used statistical techniques for external validation. The differences in the two validating approaches are highlighted, and evidence is presented that only models that have been validated externally, after their internal validation, can be considered reliable and applicable for both external prediction and regulatory purposes.

Từ khóa


Tài liệu tham khảo

Regulation (EC) No 1907/2006 of the European Parliament and of the Council (18/12/2006) concerning REACH; http://eur‐lex.europa.eu /LexUriServ/site/en/oj/2006/l_396/l_39620061230en00010849.pdf (accessed January 22 2007).

http://ecb.jrc.it/qsar/ (accessed January 22 2007).

10.1289/ehp.5757

http://www.oecd.org/dataoecd/33/37/37849783.pdf (accessed January 22 2007).

10.1080/1062936021000043472

10.1002/qsar.200390026

10.1016/S1352-2310(03)00293-0

10.1002/qsar.200390027

10.1080/1062936032000101484

10.1021/ci049923u

10.1016/j.atmosenv.2004.07.026

10.1021/ci050212l

10.1002/qsar.200530123

10.1002/qsar.200390007

P. Gramatica Evaluation of different statistical approaches to the validation of Quantitative Structure–Activity Relationships 2004.Available online at: http://ecb.jrc.it/DOCUMENTS/QSAR/Report_on_QSAR_validation_methods.pdf (accessed January 22 2007).

European Commission – QSAR for Predicting Fate and Effects of Chemicals in the Environment Final report of DG XII contract No. EV5V‐CT92‐0211 1995.

10.1021/es022457

10.1016/S0045-6535(03)00609-X

10.1080/1062936032000101510

Netzeva T. I., 2005, ATLA, 33, 155

Atkinson A. C., 1985, Plots, Transformations and Regression

10.1289/ehp.5758

10.1021/ci0342472

Jaworska J., 2005, ATLA, 33, 445

10.1021/es049220t

10.1002/cem.1180060506

10.1016/S0169-7439(00)00102-7

10.1080/01621459.1993.10476299

10.1016/S1093-3263(01)00123-1

10.1002/1521-3838(200210)21:4<348::AID-QSAR348>3.0.CO;2-D

10.1016/S0165-9936(03)00607-1

10.1007/s10822-004-4071-5

10.1023/A:1020869118689

10.1023/A:1025386326946

10.1016/j.jmgm.2004.03.003

10.1016/j.jmgm.2005.03.003

10.1897/01-171

Tong W., 2004, Predicting Chemical Toxicity and Fate, 285

10.1021/ci0498766

10.1021/tx0498253

10.1016/j.atmosenv.2005.01.007

10.1021/ci034046y

10.1021/ci0001637

10.1021/ci025626i

Kraker J. J., 2006, Chemom. Intell. Lab. Syst.

10.1080/10629360412331297407

10.1080/1062936032000169642

ENV/JM/MONO (2004) 24. Available online at: http://appli1.oecd.org/olis/2004doc.nsf/linkto/env‐jm‐mono (2004) 24 p. 122 (accessed January 22 2007).

10.1080/10629360601054388

10.1016/j.jmgm.2006.06.005

Sjostrom M., 1995, Chemometric Methods in Molecular Design

10.1016/0169-7439(92)80076-G

10.1002/anie.199305031

10.1016/S0169-7439(97)00030-0

10.1080/10629360412331297489

Schultz T. W., 2004, Predicting Chemical Toxicity and Fate, 265

10.1002/9783527615452.ch5

Thông tin
Thông tin xuất bản

Wiley

Tập 26 Số 5

694-701

Thông tin tác giả