SummaryChromatin distribution re¯ects the organization of the DNA of a nucleus and contains important cellular diagnostic and prognostic information. Feulgen staining of breast tissue enables the chromatin distribution of the nucleus to be visualized in the form of texture. Describing texture in an objective and quantitative way by means of a set of texture parameters, combined with the study of the relationship of such parameters to the pathobiological cell properties, is useful both for reduction of the subjectivity inherently coupled to visual observation and for more accurate prognosis or diagnosis.We have presented an automated classi®cation scheme for the diagnosis and grading of invasive breast cancer. The input to this scheme was a digitized microscopical image, from which nuclei were segmented. Chromatin texture was described using a set of textural parameters that include ®rst-and second-order statistics of the image grey levels. The more recently developed wavelet energy parameters were also included in our study. Classi®cation was performed by a Knn-classi®er, which is a versatile multivariate statistical classi®cation technique.We investigated the role of the tissue preparation technique and found that parameters derived from cytospins were better texture descriptors than those from sections. A 100% correct classi®cation was achieved in a patient diagnosis experiment and 82% in a nuclear grading experiment.

The influence of fixation, air-drying and liquid-imaging on the morphology as well as on the viscoelasticity of malignant mesothelioma cells was studied by atomic force microscopy. In this study, dehydrated cells were more easily scanned and offered faster data recording than hydrated cells. However, the influence of fixation strength was more noticeable. Strong fixation induced flattening of the cytoplasm and loss of nuclear structure, resulting in a clearly visible cytoskeleton which could be easily seen as fibres orientated in the direction of the cell growth. By contrast, the morphology of hydrated cells was influenced to a lesser degree on fixation and showed an overall 'rounding' of the surface with vague, ill-defined structures. Nuclear areas of these samples were difficult to image. Viscoelasticity measurements also exhibited large differences. Dehydrated cells were much harder and showed a uniform indentation profile over the whole cell that was independent of fixation. Indentation on hydrated cells was large and depended on the height of the measuring spot, the submembranous structure and, to a lesser extent, on fixation. To calculate an overall 'cellular' viscoelasticity, different methods were tested on these samples. Indentations of multiple, randomly chosen points, covering the whole cell, were measured and averaged to yield a mean indentation score. We avoided the thin and shadowed areas since it was shown that these regions were less suited for measuring. Using this design, large viscoelasticity differences were found, on which the influence of the external parameters could be shown. In another set-up, layered imaging was tried. However, long data acquisition times caused cellular activation and rearrangement, making this scanning mode unsatisfactory.

No abstract

We study ray dynamics inside the Hyperlens, a device recently demonstrated as capable of sub-diffraction-limited far-field imaging.The obtained semiclassical result of spiraling rays is confirmed by numerical simulations of gaussian beam scattering from the hyperlens.A conventional lens cannot construct the image of an object with resolution better than λ/2, where λ is the wavelength of the illuminating wave. This is because the conventional lens brings to focus only the propagating waves emanating from the object while the evanescent waves decay away exponentially and are lost in the far field. Such an apparent loss of information gives rise to the diffraction limit of λ/2. It is therefore highly desirable for many practical applications, such as biological imaging, to be able to construct images with feature sizes much below the wavelength of the illuminating light [1].Recently it was proposed that a hollow core cylinder or half-cylinder made of materials with a strongly anisotropic dielectric response [2,3] can function as a far-field imaging device capable of resolution beyond the diffraction limit [4]. The object is placed inside the hollow core and the magnified image is projected into the farfield which can be processed by conventional optics. The local wavelength at the core of the device, where the object to be imaged is placed, is below the free space wavelength and this leads to the subwavelength resolution [4,5]. The hyperbolic dispersion in strongly anisotropic materials (materials with dielectric permitivitties of opposite signs in two perpendicular directions) is the key to achieving the wavelength compression in the Optical Hyperlens.The advantages of semiclassical approach are two-fold. First, the connection to the underlying ray optics uncovers the physical origin of light propagation and imaging in the device. Second, as opposed to 'brute-force' numerical methods that are computationally intensive and suffer from instabilities when treating evanescent fields, the semiclassical approach while quantitatively accurate, is both numerically inexpensive and stable.The accuracy of the semiclassical approach in the hyperlens is due to the wavelength compression in this device. As the light approaches the core of the hyperlens, due to the hyperbolic dispersion relation the radial and tangential momentum increase, leading to substantial suppression of wavelength. Fig. 1 (a) Schematic of the hyperlens (hollow inner core) with alternating layers of metal and dielectric to achieve ε r <0 and ε θ >0 (b) path of 2 rays (shown in red) with different impact parameters impinging from vacuum onto a homogeneous medium with cylindrical anisotropy ε r = -1 and ε θ = 1 calculated using eq. (1) (c) path of the ray calculated using eq. (1) for ε r = -0.01 and ε θ = 1. Note that the rays spiral towards the center. a2032_1.pdf JWA2.pdf ©OSA 1-55752-834-9

Ensemble simulations are a promising technique for identifying the signal of atmospheric response to extra-tropical sea surface temperature variability with high statistical significance. The basic idea is to perform multiple simulations from slightly different initial conditions and then to study the average signal of the ensemble. A significant obstacle to performing such ensemble simulations is the bookkeeping required to prepare, execute, and track the progress of hundreds of different computations. We describe an ensemble simulation experiment in which the Fast Ocean Atmosphere Model was run on the U.S. TeraGrid. In this experiment, we used the GriPhyN Virtual Data System to manage our ensemble simulations and their execution on distributed resources, achieving dramatic (order-of-magnitude) reductions in turnaround time relative to previous manual experiments.