Nano‐SAR Development for Bioactivity of Nanoparticles with Considerations of Decision Boundaries

• DOI: 10.1002/smll.201201903
• Publication Date: 2013-02-19T19:20:22Z
• Authors: Rong Liu; Robert Rallo; Ralph Weissleder; Carlos Tassa; Stanley Shaw; Yoram Cohen
• Keywords: Structure-Activity Relationship; Nanoparticles; Bayes Theorem; 02 engineering and technology; 0210 nano-technology

The development of classification nano-structure-activity Relationships (nano-SARs) nanoparticle (NP) bioactivity is presented with the aim demonstrating integration multiparametric toxicity/bioactivity assays to arrive at statistically meaningful class definitions (i.e., bioactivity/inactivity endpoints), as well implications nano-SAR applicability domains and decision boundaries. Nano-SARs are constructed based on a dataset 44 iron oxide core nanoparticles (NPs), used in molecular imaging nano-sensing, containing profiles for four cell types different assays. Class developed basis 'hit' significant bioactivity) identification analysis self-organizing map consensus clustering; these enable construction nano-SARs high accuracy (>78%) NP descriptor combinations that include primary size, spin-lattice spin-spin relaxivities, zeta potentials. Analysis performance suggests H4 least hits) reasonable endpoint (from 'regulatory' viewpoint) keeping level false negatives incorrect labeling bioactive NPs inactive) low. establishment quantitative domain demonstrated, making use probability density definition naive Bayesian classifier (NBC) model (with relaxivity potential descriptors). Decision boundaries determined above H4/NBC acceptance levels negative positive predictions, illustrating practical approach may assist regulatory consideration reducing likelihood identifying being inactive.