Evaluation criteria for the quality of published experimental data on nanomaterials and their usefulness for QSAR modelling.

Nowadays nanotechnology is one of the most promising areas of science. The number and quantity of synthesized nanomaterials increase exponentially, therefore it is reasonable to expect that comprehensive risk assessment based only on empirical testing of all novel engineered nanoparticles (NPs) will very soon become impossible. Hence, the development of computational methods complementary to experimentation is very important. Quantitative structure-property relationship (QSPR) and quantitative structure-activity relationship (QSAR) models widely used in pharmaceutical chemistry and environmental science can also be modified and adopted for nanotechnology to predict physico-chemical properties and toxicity of empirically untested nanomaterials. All QSPR/QSAR modelling activities are based on experimentally derived data. It is important that, within a given data set, all values should be consistent, of high quality and measured according to a standardized protocol. Unfortunately, the amount of such data available for engineered nanoparticles in various data sources (i.e. databases and the literature) is very limited and seldom measured with a standardized protocol. Therefore, we have proposed a framework for collecting and evaluating the existing data, with the focus on possible applications for computational evaluation of properties and biological activities of nanomaterials.

Reconstructive elasticity imaging for large deformations.

A method is presented to reconstruct the elastic modulus of soft tissue based on ultrasonic displacement and strain images for comparatively large deformations. If the average deformation is too large to be described with a linear elastic model, nonlinear displacement-strain relations must be used and the mechanical equilibrium equations must include high order spatial derivatives of the displacement. Numerical methods were developed to reduce error propagation in reconstruction algorithms, including these higher order derivatives. Problems arising with the methods, as well as results using ultrasound measurements on gel-based, tissue equivalent phantoms, are given. Comparison to reconstructions using a linear elastic model shows that equivalent image quality can be produced with algorithms appropriate for finite amplitude deformations.