Humans and most animals can learn new tasks without forgetting old ones. However, training artificial neural networks (ANNs) on new tasks typically causes them to forget previously learned tasks. ...This phenomenon is the result of “catastrophic forgetting,” in which training an ANN disrupts connection weights that were important for solving previous tasks, degrading task performance. Several recent studies have proposed methods to stabilize connection weights of ANNs that are deemed most important for solving a task, which helps alleviate catastrophic forgetting. Here, drawing inspiration from algorithms that are believed to be implemented in vivo, we propose a complementary method: adding a context-dependent gating signal, such that only sparse, mostly nonoverlapping patterns of units are active for any one task. This method is easy to implement, requires little computational overhead, and allows ANNs to maintain high performance across large numbers of sequentially presented tasks, particularly when combined with weight stabilization. We show that this method works for both feedforward and recurrent network architectures, trained using either supervised or reinforcement-based learning. This suggests that using multiple, complementary methods, akin to what is believed to occur in the brain, can be a highly effective strategy to support continual learning.
Climate strongly influences global wildfire activity, and recent wildfire surges may signal fire weather-induced pyrogeographic shifts. Here we use three daily global climate data sets and three fire ...danger indices to develop a simple annual metric of fire weather season length, and map spatio-temporal trends from 1979 to 2013. We show that fire weather seasons have lengthened across 29.6 million km(2) (25.3%) of the Earth's vegetated surface, resulting in an 18.7% increase in global mean fire weather season length. We also show a doubling (108.1% increase) of global burnable area affected by long fire weather seasons (>1.0 σ above the historical mean) and an increased global frequency of long fire weather seasons across 62.4 million km(2) (53.4%) during the second half of the study period. If these fire weather changes are coupled with ignition sources and available fuel, they could markedly impact global ecosystems, societies, economies and climate.
There is mounting concern that global wildfire activity is shifting in frequency, intensity, and seasonality in response to climate change. Fuel moisture provides a powerful means of detecting ...changing fire potential. Here, we use global burned area, weather reanalysis data, and the Canadian fire weather index system to calculate fuel moisture trends for multiscale biogeographic regions across a gradient in vegetation productivity. We quantify the proportion of days in the local fire season between 1979 and 2019, where fuel moisture content is below a critical threshold indicating extreme fire potential. We then associate fuel moisture trends over that period to vegetation productivity and comment on its implications for projected anthropogenic climate change. Overall, there is a strong drying trend across realms, biomes, and the productivity gradient. Even where a wetting trend is observed, this often indicates a trend toward increasing fire activity due to an expected increase in fuel production. The detected trends across the productivity gradient lead us to conclude global fire activity will increase with anthropogenic climate change.
The last 40 years have seen a strong drying trend across wildfire fuels worldwide. Even where a wetting trend is observed, this often indicates a trend toward increasing fire activity due to an expected increase in fuel production. These detected trends lead us to conclude global fire activity will increase with anthropogenic climate change.
It is currently unclear whether the GluN2 subtype influences NMDA receptor (NMDAR) excitotoxicity. We report that the toxicity of NMDAR-mediated Ca2+ influx is differentially controlled by the ...cytoplasmic C-terminal domains of GluN2B (CTD2B) and GluN2A (CTD2A). Studying the effects of acute expression of GluN2A/2B-based chimeric subunits with reciprocal exchanges of their CTDs revealed that CTD2B enhances NMDAR toxicity, compared to CTD2A. Furthermore, the vulnerability of forebrain neurons in vitro and in vivo to NMDAR-dependent Ca2+ influx is lowered by replacing the CTD of GluN2B with that of GluN2A by targeted exon exchange in a mouse knockin model. Mechanistically, CTD2B exhibits stronger physical/functional coupling to the PSD-95-nNOS pathway, which suppresses protective CREB activation. Dependence of NMDAR excitotoxicity on the GluN2 CTD subtype can be overcome by inducing high levels of NMDAR activity. Thus, the identity (2A versus 2B) of the GluN2 CTD controls the toxicity dose-response to episodes of NMDAR activity.
► The CTD of GluN2B promotes excitotoxicity better than that of GluN2A ► GluN2 CTD subtype differences are seen in both WT and chimeric 2A/2B subunits ► The GluN2B CTD couples to a prodeath PSD-95/nNOS-dependent CREB shut-off pathway
Martel et al. find that the two subtypes (2A versus 2B) of the GluN2 C-terminal domain differentially couple to the CREB shut-off pathway, causing distinct effects on NMDA receptor-mediated neuronal death both in vitro and in vivo.
Most marketed pharmaceuticals consist of molecular crystals. The arrangement of the molecules in a crystal determines its physical properties and, in certain cases, its chemical properties, and so ...greatly influences the processing and formulation of solid pharmaceuticals, as well as key drug properties such as dissolution rate and stability. A thorough understanding of the relationships between physical structures and the properties of pharmaceutical solids is therefore important in selecting the most suitable form of an active pharmaceutical ingredient for development into a drug product. In this article, we review the different crystal forms of pharmaceuticals, the challenges that they present and recent advances in crystal structure determination. We then discuss computational approaches for predicting crystal properties. Finally, we review the analysis of crystal structures in furthering crystal engineering to design novel pharmaceutical compounds with desired physical and mechanical properties.
The ancient UbiB protein kinase-like family is involved in isoprenoid lipid biosynthesis and is implicated in human diseases, but demonstration of UbiB kinase activity has remained elusive for ...unknown reasons. Here, we quantitatively define UbiB-specific sequence motifs and reveal their positions within the crystal structure of a UbiB protein, ADCK3. We find that multiple UbiB-specific features are poised to inhibit protein kinase activity, including an N-terminal domain that occupies the typical substrate binding pocket and a unique A-rich loop that limits ATP binding by establishing an unusual selectivity for ADP. A single alanine-to-glycine mutation of this loop flips this coenzyme selectivity and enables autophosphorylation but inhibits coenzyme Q biosynthesis in vivo, demonstrating functional relevance for this unique feature. Our work provides mechanistic insight into UbiB enzyme activity and establishes a molecular foundation for further investigation of how UbiB family proteins affect diseases and diverse biological pathways.
Display omitted
•A UbiB family protein crystal structure reveals an atypical kinase-like fold•An A-rich loop determines coenzyme selectivity and inhibits autophosphorylation•UbiB-specific features are required for coenzyme Q biosynthesis in vivo•Pathogenic ADCK3 and ADCK4 mutations disrupt protein stability
The largely uncharacterized UbiB protein kinase-like family is conserved across all superkingdoms of life. Stefely et al. find that ADCK3, a mitochondrial UbiB protein, adopts an atypical structure with unique features positioned to inhibit protein kinase activity. Altering these features enables in vitro autophosphorylation, but compromises in vivo coenzyme Q biosynthesis.
Topology optimization has evolved rapidly since the late 1980s. The optimization of the geometry and topology of structures has a great impact on its performance, and the last two decades have seen ...an exponential increase in publications on structural optimization. This has mainly been due to the success of material distribution methods, originating in 1988, for generating optimal topologies of structural elements. Previous methods suffered from mathematical complexity and a limited scope for applicability, however with the advent of increased computational power and new techniques topology optimization has grown into a design tool used by industry. There are two main fields in structural topology optimization, gradient based, where mathematical models are derived to calculate the sensitivities of the design variables, and non gradient based, where material is removed or included using a sensitivity function. Both fields have been researched in great detail over the last two decades, to the point where structural topology optimization has been applied to real world structures. It is the objective of this review paper to present an overview of the developments in non gradient based structural topology and shape optimization, with a focus on evolutionary algorithms, which began as a non gradient method, but have developed to incorporate gradient based techniques. Starting with the early work and development of the popular algorithms and focusing on the various applications. The sensitivity functions for various optimization tasks are presented and real world applications are analyzed. The article concludes with new applications of topology optimization and applications in various engineering fields.
Ever-harsher service conditions in the future will call for materials with increasing ability to undergo deformation without sustaining damage while retaining high strength. Prime candidates for ...these conditions are certain high-entropy alloys (HEAs), which have extraordinary work-hardening ability and toughness. By subjecting the equiatomic CrMnFeCoNi HEA to severe plastic deformation through swaging followed by either quasi-static compression or dynamic deformation in shear, we observe a dense structure comprising stacking faults, twins, transformation from the face-centered cubic to the hexagonal close-packed structure, and, of particular note, amorphization. The coordinated propagation of stacking faults and twins along {111} planes generates high-deformation regions, which can reorganize into hexagonal packets; when the defect density in these regions reaches a critical level, they generate islands of amorphous material. These regions can have outstanding mechanical properties, which provide additional strengthening and/or toughening mechanisms to enhance the capability of these alloys to withstand extreme loading conditions.
Intraocular pressure (IOP) is currently the sole modifiable risk factor for primary open-angle glaucoma (POAG), one of the leading causes of blindness worldwide
. Both IOP and POAG are highly ...heritable
. We report a combined analysis of participants from the UK Biobank (n = 103,914) and previously published data from the International Glaucoma Genetic Consortium (n = 29,578)
that identified 101 statistically independent genome-wide-significant SNPs for IOP, 85 of which have not been previously reported
. We examined these SNPs in 11,018 glaucoma cases and 126,069 controls, and 53 SNPs showed evidence of association. Gene-based tests implicated an additional 22 independent genes associated with IOP. We derived an allele score based on the IOP loci and loci influencing optic nerve head morphology. In 1,734 people with advanced glaucoma and 2,938 controls, participants in the top decile of the allele score were at increased risk (odds ratio (OR) = 5.6; 95% confidence interval (CI): 4.1-7.6) of glaucoma relative to the bottom decile.
Covalent organic frameworks (COFs) are porous crystalline materials composed of light elements linked by strong covalent bonds. A number of these materials contain a high density of Lewis acid boron ...sites that can strongly interact with Lewis basic guests, which makes them ideal for the storage of corrosive chemicals such as ammonia. We found that a member of the covalent organic framework family, COF-10, shows the highest uptake capacity (15 mol kg⁻¹, 298 K, 1 bar) of any porous material, including microporous 13X zeolite (9 mol kg⁻¹), Amberlyst 15 (11 mol kg⁻¹) and mesoporous silica, MCM-41 (7.9 mol kg⁻¹). Notably, ammonia can be removed from the pores of COF-10 by heating samples at 200°C under vacuum. In addition, repeated adsorption of ammonia into COF-10 causes a shift in the interlayer packing, which reduces its apparent surface area to nitrogen. However, owing to the strong Lewis acid-base interactions, the total uptake capacity of ammonia and the structural integrity of the COF are maintained after several cycles of adsorption/desorption.