In this study, we investigated the formation of radiation-induced foci in normal human fibroblasts exposed to X rays or 130 keV/mum nitrogen ions using antibodies to phosphorylated protein kinase ataxia telangiectasia mutated (ATMp) and histone H2AX (gamma-H2AX). High-content automatic image analysis was used to quantify the immunofluorescence of radiation-induced foci. The size of radiation-induced foci increased for both proteins over a 2-h period after nitrogen-ion irradiation, while the size of radiation-induced foci did not change after exposure to low-LET radiation. The number of radiation-induced ATMp foci showed a more rapid rise and greater frequency after X-ray exposure and was resolved more rapidly such that the frequency of radiation-induced foci decreased by 90% compared to 60% after exposure to high-LET radiation 2 h after 30 cGy. In contrast, the kinetics of radiation-induced gamma-H2AX focus formation was similar for high- and low-LET radiation in that it reached a plateau early and remained constant for up to 2 h. High-resolution 3D images of radiation-induced gamma-H2AX foci and dosimetry computation suggest that multiple double-strand breaks from nitrogen ions are encompassed within large nuclear domains of 4.4 Mbp. Our work shows that the size and frequency of radiation-induced foci vary as a function of radiation quality, dose, time and protein target. Thus, even though double-strand breaks and radiation-induced foci are correlated, the dynamic nature of both contradicts their accepted equivalence for low doses of different radiation qualities.

We present the results of applying new object classification techniques to difference images in the context of the Nearby Supernova Factory supernova search. Most current supernova searches subtract reference images from new images, identify objects in these difference images, and apply simple threshold cuts on parameters such as statistical significance, shape, and motion to reject objects such as cosmic rays, asteroids, and subtraction artifacts. Although most static objects subtract cleanly, even a very low false positive detection rate can lead to hundreds of non-supernova candidates which must be vetted by human inspection before triggering additional followup. In comparison to simple threshold cuts, more sophisticated methods such as Boosted Decision Trees, Random Forests, and Support Vector Machines provide dramatically better object discrimination. At the Nearby Supernova Factory, we reduced the number of non-supernova candidates by a factor of 10 while increasing our supernova identification efficiency. Methods such as these will be crucial for maintaining a reasonable false positive rate in the automated transient alert pipelines of upcoming projects such as PanSTARRS and LSST.

We present a set of high-resolution global atmospheric general circulation model (AGCM) simulations focusing on the model's ability to represent tropical storms and their statistics. We find that the model produces storms of hurricane strength with realistic dynamical features. We also find that tropical storm statistics are reasonable, both globally and in the north Atlantic, when compared to recent observations. The sensitivity of simulated tropical storm statistics to increases in sea surface temperature (SST) is also investigated, revealing that a credible late 21st century SST increase produced increases in simulated tropical storm numbers and intensities in all ocean basins. While this paper supports previous high-resolution model and theoretical findings that the frequency of very intense storms will increase in a warmer climate, it differs notably from previous medium and high-resolution model studies that show a global reduction in total tropical storm frequency. However, we are quick to point out that this particular model finding remains speculative due to a lack of radiative forcing changes in our time-slice experiments as well as a focus on the Northern hemisphere tropical storm seasons.

We introduce a novel application of Support Vector Machines (SVMs) to the problem of identifying potential supernovae using photometric and geometric features computed from astronomical imagery. The challenges of this supervised learning application are significant: 1) noisy and corrupt imagery resulting in high levels of feature uncertainty, 2) features with heavy-tailed, peaked distributions, 3) extremely imbalanced and overlapping positive and negative data sets, and 4) the need to reach high positive classification rates, i.e. to find all potential supernovae, while reducing the burdensome workload of manually examining false positives. High accuracy is achieved via a sign-preserving, shifted log transform applied to features with peaked, heavy-tailed distributions. The imbalanced data problem is handled by oversampling positive examples, selectively sampling misclassified negative examples, and iteratively training multiple SVMs for improved supernova recognition on unseen test data. We present crossvalidation results and demonstrate the impact on a largescale supernova survey that currently uses the SVM decision value to rank-order 600,000 potential supernovae each night.