Explore the vast range of titles available.

Abnormal grain growth (AGG) in polycrystalline microstructures, characterized by the rapid and disproportionate enlargement of a few “abnormal” grains relative to their surroundings, can lead to dramatic, often deleterious changes in the mechanical properties of materials, such as strength and toughness. Thus, the prediction and control of AGG is key to realizing robust mesoscale materials design. Unfortunately, it is challenging to predict these rare events far in advance of their onset because, at early stages, there is little to distinguish incipient abnormal grains from “normal” grains. To overcome this difficulty, we propose two machine learning approaches for predicting whether a grain will become abnormal in the future. These methods analyze grain properties derived from the spatio-temporal evolution of grain characteristics, grain-grain interactions, and a network-based analysis of these relationships. The first, PAL ( P redicting A bnormality with L STM), analyzes grain features using a long short-term memory (LSTM) network, and the second, PAGL ( P redicting A bnormality with G CRN and L STM), supplements the LSTM with a graph-based convolutional recurrent network (GCRN). We validated these methods on three distinct material scenarios with differing grain properties, observing that PAL and PAGL achieve high sensitivity and precision and, critically, that they are able to predict future abnormality long before it occurs. Finally, we consider the application of the deep learning models developed here to the prediction of rare events in different contexts.

Due to the difficulty in observing grain growth in real samples, grain growth simulations have become commonplace for analyzing and predicting grain behavior in manufacturing processes. In this work, we describe the theoretical basis for these simulations and their mechanisms, as well as how they are commonly implemented. We provide a novel transformed boundary propagation mechanism as well as novel simulation optimizations, such as using regional octrees for tracking subsystem activity, tracking grain boundaries sparsely in real-time, and simulating grain growth via parallel subregion simulation. We also provide information on how to properly store and visualize simulation outputs, as well as how to use those outputs for predictive machine learning tasks.

The accurate prediction and estimation of annual snow accumulation has grown in importance as we deal with the effects of climate change and the increase of global atmospheric temperatures. Airborne radar sensors, such as the Snow Radar, are able to measure accumulation rate patterns at a large-scale and monitor the effects of ongoing climate change on Greenland's precipitation and run-off. The Snow Radar's use of an ultra-wide bandwidth enables a fine vertical resolution that helps in capturing internal ice layers. Given the amount of snow accumulation in previous years using the radar data, in this paper, we propose a machine learning model based on recurrent graph convolutional networks to predict the snow accumulation in recent consecutive years at a certain location. We found that the model performs better and with more consistency than equivalent nongeometric and nontemporal models.

As we deal with the effects of climate change and the increase of global atmospheric temperatures, the accurate tracking and prediction of ice layers within polar ice sheets grows in importance. Studying these ice layers reveals climate trends, how snowfall has changed over time, and the trajectory of future climate and precipitation. In this paper, we propose a machine learning model that uses adaptive, recurrent graph convolutional networks to, when given the amount of snow accumulation in recent years gathered through airborne radar data, predict historic snow accumulation by way of the thickness of deep ice layers. We found that our model performs better and with greater consistency than our previous model as well as equivalent non-temporal, non-geometric, and non-adaptive models.

The precise tracking and prediction of polar ice layers can unveil historic trends in snow accumulation. In recent years, airborne radar sensors, such as the Snow Radar, have been shown to be able to measure these internal ice layers over large areas with a fine vertical resolution. In our previous work, we found that temporal graph convolutional networks perform reasonably well in predicting future snow accumulation when given temporal graphs containing deep ice layer thickness. In this work, we experiment with a graph attention network-based model and used it to predict more annual snow accumulation data points with fewer input data points on a larger dataset. We found that these large changes only very slightly negatively impacted performance.

In bacterial systems, CRISPR-Cas transcriptional activation (CRISPRa) has the potential to dramatically expand our ability to regulate gene expression, but we lack predictive rules for designing effective gRNA target sites. Here, we identify multiple features of bacterial promoters that impose stringent requirements on CRISPRa target sites. Notably, we observe narrow, 2–4 base windows of effective sites with a periodicity corresponding to one helical turn of DNA, spanning ~40 bases and centered ~80 bases upstream of the TSS. However, we also identify two features suggesting the potential for broad scope: CRISPRa is effective at a broad range of σ 70 -family promoters, and an expanded PAM dCas9 allows the activation of promoters that cannot be activated by S. pyogenes dCas9. These results provide a roadmap for future engineering efforts to further expand and generalize the scope of bacterial CRISPRa. The use of CRISPRa in bacteria lacks predictive rules for identifying effective gRNA target sites. Here the authors define features of bacterial promoters that impose stringent requirements on effective sites.

Abstract In bacterial systems, CRISPR-Cas transcriptional activation (CRISPRa) has the potential to dramatically expand our ability to regulate gene expression, but we currently lack a complete understanding of the rules for designing effective guide RNA target sites. We have identified multiple features of bacterial promoters that impose stringent requirements on bacterial CRISPRa target sites. Most importantly, we found that shifting a gRNA target site by 2-4 bases along the DNA target can cause a nearly complete loss in activity. The loss in activity can be rescued by shifting the target site 10-11 bases, corresponding to one full helical turn. Practically, our results suggest that it will be challenging to find a gRNA target site with an appropriate PAM sequence at precisely the right position at arbitrary genes of interest. To overcome this limitation, we demonstrate that a dCas9 variant with expanded PAM specificity allows activation of promoters that cannot be activated by S. pyogenes dCas9. These results provide a roadmap for future engineering efforts to further expand and generalize the scope of bacterial CRISPRa.

CRISPR-Cas transcriptional tools have been widely applied for programmable regulation of complex biological networks. In comparison to eukaryotic systems, bacterial CRISPR activation (CRISPRa) has stringent target site requirements for effective gene activation. While genes may not always have an NGG PAM at the appropriate position, PAM-flexible dCas9 variants can expand the range of targetable sites. Here we systematically evaluate a panel of PAM-flexible dCas9 variants for their ability to activate bacterial genes. We observe that dxCas9-NG provides a high dynamic range of gene activation for sites with NGN PAMs while dSpRY permits modest activity across almost any PAM. Similar trends were observed for heterologous and endogenous promoters. For all variants tested, improved PAM-flexibility comes with the tradeoff that CRISPRi-mediated gene repression becomes less effective. Weaker CRISPRi gene repression can be partially rescued by expressing multiple sgRNAs to target many sites in the gene of interest. Our work provides a framework to choose the most effective dCas9 variant for a given set of gene targets, which will further expand the utility of CRISPRa/i gene regulation in bacterial systems.