Nanoparticles (NPs) are increasingly used throughout the world for many purposes. The resulting exposure increases the relevance of efforts to assess their effects. The activities of NPs are related to many structural features, including their shape, composition and size. Applying Quantitative Structure–Activity Relationship (QSAR) methods to nanoscale systems becomes challenging, due to the lack of data and insight into the fate and effects of NPs. In this study, the possible use of QSAR methods on NPs is investigated. To this intent, several ways of representing and describing NPs were tested by using different data mining methods. The main conclusion is that QSAR methods are relevant for the study of the activity of NPs, but this should be confirmed by using larger and more diverse sets of data. Moreover, representing the constitution of NPs (in terms of core, coating and surface modification) significantly increases the prediction accuracy of the models. In our case, the most significant features to be represented were found to be the core and surface modification.
Anon. (2012). ASTM Standard E 2456-06(2012), 4pp. West Conshohocken, PA, USA: ASTM International.
2.
ShevchenkoV.Y., MadisonA.E. & MackayA.L. (2003). The structural diversity of the nanoworld. Glass Physics & Chemistry296, 577–582.
3.
MaynardA.D. & AitkenR.J. (2007). Assessing exposure to airborne nanomaterials: Current abilities & future requirements. Nanotoxicology11, 26–41.
4.
OberdörsterG., MaynardA., DonaldsonK., CastranovaV., FitzpatrickJ., AusmanK., CarterJ., KarnB., KreylingW., LaiD., OlinS., Monteiro-RiviereN., WarheitD. & YangH. (2005). Principles for characterizing the potential human health effects from exposure to nanomaterials: Elements of a screening strategy. Particle & Fibre Toxicology28.
5.
TodeschiniR. & ConsoniV. (2003). Handbook of Molecular Descriptors, Volume 3, pp. 12. Weinheim, Germany: Wiley-VCH.
6.
LivingstoneD.J. (ed.) (2008). Artificial Neural Networks: Methods & Applications, pp. 185–202. New York, NY, USA: Humana Press.
7.
AjayA. (1993). A unified framework for using neural networks to build QSARs. Journal of Medicinal Chemistry36, 3565–3571.
WoldH. (1982). Soft modeling: The basic design and some extensions. In Systems Under Indirect Observation, Vol. 2 (ed. JöreskogK.G. & WoldH.) pp. 1–54. Amsterdam, The Netherlands: Elsevier Science Ltd.
10.
WoldS. (1995). PLS for multivariate linear modeling. In Chemometrics Methods in Molecular Design, pp. 195–218. Weinheim, Germany: Wiley-VCH.
11.
ClementiS. & WoldS. (1995). How to choose the proper statistical method. In Chemometrics Methods in Molecular Design. pp. 319–336. Weinheim, Germany: Wiley-VCH.
12.
GolbraikhA. & TropshaA. (2002). Beware of q2!Journal of Molecular Graphics & Modelling20, 269–276.
13.
NettletonD.F., Orriols-PuigA. & FornellsA. (2010). A study of the effect of different types of noise on the precision of supervised learning technique. Artificial Intelligence Review33, 275–306.
14.
HawkinsD.M. (2004). The problem of overfitting. Journal of Chemical Information & Computer Sciences44, 1–12.
15.
ThomasD.G., PappuR.V. & BakerN.A. (2011). Nanoparticle ontology for cancer nanotechnology research. Journal of Biomedical Informatics441, 59–74.
RSC (2013). ChemSpider. Cambridge, UK: Royal Society of Chemistry. Available at: http://www.chemspider.com/ (Accessed 20.12.13).
21.
MauriA., ConsonniV., PavanM. & TodeschiniR. (2006). DRAGON software: An easy approach to molecular descriptor calculations. MATCH Communications in Mathematical & in Computer Chemistry562, 237–248.
22.
YapC.W. (2011). PaDEL-descriptor: An open source software to calculate molecular descriptors and fingerprints. Journal of Computational Chemistry, 327, 1466–1474.
23.
HallL.H. & KierL.B. (1995). Electrotopological state indices for atom types: A novel combination of electronic, topological, and valence state information. Journal of Chemical Information & Computer Sciences356, 1039–1045.
24.
HallL.H., KierL.B. & BrownB.B. (1995). Molecular similarity based on novel atom-type electrotopological state indices. Journal of Chemical Information & Computer Sciences356, 1074–1080.
25.
SteinbeckC., HanY., KuhnS., HorlacherO., LuttmannE. & WillighagenE.L. (2003). The chemistry development kit (CDK): An open-source java library for chemo- and bioinformatics. Journal of Chemical Information & Computer Sciences, 432, 493–500.
26.
SteinbeckC., HoppeC., KuhnS., FlorisM., GuhaR. & WillighagenE.L. (2006). Recent developments of the chemistry development kit (CDK): An open-source java library for chemo- and bioinformatics. Journal of Chemical Information & Computer Sciences1217, 2111–2120.
27.
HallM., FrankE., HolmesG., PfahringerB., ReutemannP. & WittenI.H. (2009). The WEKA data mining software: An update. SIGKDD Explorations11, 10–18.
28.
IhakaR. & GentlemanR. (1996). R: A language for data analysis & graphics. Journal of Computational & Graphical Statistics53, 299–314.
