Explore the vast range of titles available.

Improving the efficiency of algorithms for fundamental computations can have a widespread impact, as it can affect the overall speed of a large amount of computations. Matrix multiplication is one such primitive task, occurring in many systems—from neural networks to scientific computing routines. The automatic discovery of algorithms using machine learning offers the prospect of reaching beyond human intuition and outperforming the current best human-designed algorithms. However, automating the algorithm discovery procedure is intricate, as the space of possible algorithms is enormous. Here we report a deep reinforcement learning approach based on AlphaZero 1 for discovering efficient and provably correct algorithms for the multiplication of arbitrary matrices. Our agent, AlphaTensor, is trained to play a single-player game where the objective is finding tensor decompositions within a finite factor space. AlphaTensor discovered algorithms that outperform the state-of-the-art complexity for many matrix sizes. Particularly relevant is the case of 4 × 4 matrices in a finite field, where AlphaTensor’s algorithm improves on Strassen’s two-level algorithm for the first time, to our knowledge, since its discovery 50 years ago 2 . We further showcase the flexibility of AlphaTensor through different use-cases: algorithms with state-of-the-art complexity for structured matrix multiplication and improved practical efficiency by optimizing matrix multiplication for runtime on specific hardware. Our results highlight AlphaTensor’s ability to accelerate the process of algorithmic discovery on a range of problems, and to optimize for different criteria. A reinforcement learning approach based on AlphaZero is used to discover efficient and provably correct algorithms for matrix multiplication, finding faster algorithms for a variety of matrix sizes.

The development of catalysts able to assist industrially important chemical processes is a topic of high importance. In view of the catalytic capabilities of small metal clusters, research efforts are being focused on the synthesis of novel catalysts bearing such active sites. Here we report a heterogeneous catalyst consisting of Pd 4 clusters with mixed-valence 0/+1 oxidation states, stabilized and homogeneously organized within the walls of a metal–organic framework (MOF). The resulting solid catalyst outperforms state-of-the-art metal catalysts in carbene-mediated reactions of diazoacetates, with high yields (>90%) and turnover numbers (up to 100,000). In addition, the MOF-supported Pd 4 clusters retain their catalytic activity in repeated batch and flow reactions (>20 cycles). Our findings demonstrate how this synthetic approach may now instruct the future design of heterogeneous catalysts with advantageous reaction capabilities for other important processes. Mixed-valence clusters of Pd 4 organized within a metal–organic framework exhibit robust catalytic capacities during carbene-mediated chemical reactions.

The genus Citrus , comprising some of the most widely cultivated fruit crops worldwide, includes an uncertain number of species. Here we describe ten natural citrus species, using genomic, phylogenetic and biogeographic analyses of 60 accessions representing diverse citrus germ plasms, and propose that citrus diversified during the late Miocene epoch through a rapid southeast Asian radiation that correlates with a marked weakening of the monsoons. A second radiation enabled by migration across the Wallace line gave rise to the Australian limes in the early Pliocene epoch. Further identification and analyses of hybrids and admixed genomes provides insights into the genealogy of major commercial cultivars of citrus. Among mandarins and sweet orange, we find an extensive network of relatedness that illuminates the domestication of these groups. Widespread pummelo admixture among these mandarins and its correlation with fruit size and acidity suggests a plausible role of pummelo introgression in the selection of palatable mandarins. This work provides a new evolutionary framework for the genus Citrus . The origin, evolution and domestication of Citrus and the genealogy of the most important wild and cultivated citrus varieties. When life gave us lemons Citrus fruits are one of the most cultivated crops worldwide, yet the evolutionary relationships among citrus species remain uncertain. Daniel Rokhsar, Manuel Talon and colleagues analyse the genomes of 60 accessions that represent a diverse range of citrus species, including 30 newly sequenced citrus genomes. They characterize the diversity and evolution of citrus at the species level and identify interspecific citrus hybrids and admixtures—genetic mixing between previously isolated populations—that could be the result of human activities such as migration and agriculture. The authors identify 10 progenitor species and suggest that citrus originated in southeast Asia, diversifying during the late Miocene epoch through a rapid southeast Asian radiation that correlated with a changing climate, including the weakening of the monsoons. They also find extensive relatedness among mandarins and sweet oranges, showing a complex history of admixture during the domestication of these groups.

Much recent marine microbial research has focused on sponges, but very little is known about how the sponge microbiome fits in the greater coral reef microbial metacommunity. Here, we present an extensive survey of the prokaryote communities of a wide range of biotopes from Indo-Pacific coral reef environments. We find a large variation in operational taxonomic unit (OTU) richness, with algae, chitons, stony corals and sea cucumbers housing the most diverse prokaryote communities. These biotopes share a higher percentage and number of OTUs with sediment and are particularly enriched in members of the phylum Planctomycetes. Despite having lower OTU richness, sponges share the greatest percentage (>90%) of OTUs with >100 sequences with the environment (sediment and/or seawater) although there is considerable variation among sponge species. Our results, furthermore, highlight that prokaryote microorganisms are shared among multiple coral reef biotopes, and that, although compositionally distinct, the sponge prokaryote community does not appear to be as sponge-specific as previously thought.

Background The oxidation of protein-bound methionine to form methionine sulfoxide, has traditionally been regarded as an oxidative damage. However, recent evidences support the view of this reversible reaction as a regulatory post-translational modification. The perception that methionine sulfoxidation may provide a mechanism to the redox regulation of a wide range of cellular processes, has stimulated some proteomic studies. However, these experimental approaches are expensive and time-consuming. Therefore, computational methods designed to predict methionine oxidation sites are an attractive alternative. As a first approach to this matter, we have developed models based on random forests, support vector machines and neural networks, aimed at accurate prediction of sites of methionine oxidation. Results Starting from published proteomic data regarding oxidized methionines, we created a hand-curated dataset formed by 113 unique polypeptides of known structure, containing 975 methionyl residues, 122 of which were oxidation-prone (positive dataset) and 853 were oxidation-resistant (negative dataset). We use a machine learning approach to generate predictive models from these datasets. Among the multiple features used in the classification task, some of them contributed substantially to the performance of the predictive models. Thus, (i) the solvent accessible area of the methionine residue, (ii) the number of residues between the analyzed methionine and the next methionine found towards the N-terminus and (iii) the spatial distance between the atom of sulfur from the analyzed methionine and the closest aromatic residue, were among the most relevant features. Compared to the other classifiers we also evaluated, random forests provided the best performance, with accuracy, sensitivity and specificity of 0.7468±0.0567, 0.6817±0.0982 and 0.7557±0.0721, respectively (mean ± standard deviation). Conclusions We present the first predictive models aimed to computationally detect methionine sites that may become oxidized in vivo in response to oxidative signals. These models provide insights into the structural context in which a methionine residue become either oxidation-resistant or oxidation-prone. Furthermore, these models should be useful in prioritizing methinonyl residues for further studies to determine their potential as regulatory post-translational modification sites.

All real processes, be they chemical, mechanical or electrical, are thermodynamically irreversible and therefore suffer from thermodynamic losses. Here, we report the design and operation of a chemical reactor capable of approaching thermodynamically reversible operation. The reactor was employed for hydrogen production via the water–gas shift reaction, an important route to ‘green’ hydrogen. The reactor avoids mixing reactant gases by transferring oxygen from the (oxidizing) water stream to the (reducing) carbon monoxide stream via a solid-state oxygen reservoir consisting of a perovskite phase (La 0.6 Sr 0.4 FeO 3- δ ). This reservoir is able to remain close to equilibrium with the reacting gas streams because of its variable degree of non-stoichiometry and thus develops a ‘chemical memory’ that we employ to approach reversibility. We demonstrate this memory using operando, spatially resolved, real-time, high-resolution X-ray powder diffraction on a working reactor. The design leads to a reactor unconstrained by overall chemical equilibrium limitations, which can produce essentially pure hydrogen and carbon dioxide as separate product streams. Conventional chemical reactors are subject to the equilibrium limitations imposed by the overall reaction. It has now been shown that this limitation can be overcome if reactants are fed separately to a reactor and a non-stoichiometric oxygen carrier is used to transfer both oxygen and key chemical information across a reaction cycle.

Background: Bisphenol A (BPA) is one of the highest-volume chemicals produced worldwide, and human exposure to BPA is thought to be ubiquitous. Thus, there are concerns that the amount of BPA to which humans are exposed may cause adverse health effects. Importantly, results from a large number of biomonitoring studies are at odds with the results from two toxicokinetic studies. Objective: We examined several possibilities for why biomonitoring and toxicokinetic studies could come to seemingly conflicting conclusions. Data sources: We examined > 80 published human biomonitoring studies that measured BPA concentrations in human tissues, urine, blood, and other fluids, along with two toxicokinetic studies of human BPA metabolism. Data extraction and synthesis: The > 80 biomonitoring studies examined included measurements in thousands of individuals from several different countries, and these studies overwhelmingly detected BPA in individual adults, adolescents, and children. Unconjugated BPA was routinely detected in blood (in the nanograms per milliliter range), and conjugated BPA was routinely detected in the vast majority of urine samples (also in the nanograms per milliliter range). In stark contrast, toxicokinetic studies proposed that humans are not internally exposed to BPA. Some regulatory agencies have relied solely on these toxicokinetic models in their risk assessments. Conclusions: Available data from biomonitoring studies clearly indicate that the general population is exposed to BPA and is at risk from internal exposure to unconjugated BPA. The two toxicokinetic studies that suggested human BPA exposure is negligible have significant deficiencies, are directly contradicted by hypothesis-driven studies, and are therefore not reliable for risk assessment purposes.

An update to the American Cancer Society (ACS) guideline regarding screening for the early detection of cervical precancerous lesions and cancer is presented. The guidelines are based on a systematic evidence review, contributions from 6 working groups, and a recent symposium cosponsored by the ACS, the American Society for Colposcopy and Cervical Pathology, and the American Society for Clinical Pathology, which was attended by 25 organizations. The new screening recommendations address age-appropriate screening strategies, including the use of cytology and high-risk human papillomavirus (HPV) testing, follow-up (eg, the management of screen positives and screening intervals for screen negatives) of women after screening, the age at which to exit screening, future considerations regarding HPV testing alone as a primary screening approach, and screening strategies for women vaccinated against HPV16 and HPV18 infections.

Major depressive disorder (MDD) has a prevalence of 5% in adolescents. Several studies have described the association between the inflammatory response and MDD, but little is known about the relationship between MDD and growth factors, such as IL-7, IL-9, IL-17A, VEGF, basic FGF, G-CSF, and GM-CSF. It must be appointed that there are scarce reports on growth factors in adolescents with MDD and even fewer with a clinical follow-up. In this work, we evaluated the levels of growth factors (IL-7, IL-9, IL-17A, VEGF, basic FGF, G-CSF, and GM-CSF) in MDD adolescents and the clinical follow-up during eight weeks of treatment with fluoxetine. Methods. All patients were diagnosed according to the DSM-IV-TR, and the severity of the symptoms was evaluated using the Hamilton Depression Rating Scale (HDRS). Growth factors IL-7, IL-9, IL-17A, VEGF, basic FGF, G-CSF, and GM-CSF were quantified by cytometric bead array using serum samples from 22 adolescents with MDD and 18 healthy volunteers. Results. All patients showed clinical improvement since the fourth week of pharmacological treatment according to the HDRS. Considerably higher levels of IL-7, IL-9, IL-17A, VEGF, basic FGF, G-CSF, and GM-CSF were detected in MDD adolescents as compared to healthy volunteers. A significant but temporal decrease was detected in basic FGF, G-CSF, and GM-CSF at week four of fluoxetine administration. Conclusions. To the best of our knowledge, this is the first report to show alterations in the levels of growth factors, such as IL-7, IL-9, IL-17A, VEGF, basic FGF, G-CSF, and GM-CSF in MDD adolescents during eight weeks of clinical follow-up. These disturbances might be involved in the physiopathology of MDD since such growth factors have been proven to participate in the neural development and correct functioning of the CNS; therefore, subtle alterations in it may contribute to MDD.

Evidence of the biological impacts of weak magnetic fields have been reported for more than fifty years. However, research progress on such effects has been hampered by a lack of systematics in most experiments. Efforts to increase the systematics in such cell biology experiments must include the capability of producing fields that can be automatically adjusted and that are stable throughout an experiment’s duration, usually operating inside an incubator. Here, we report on the design of a fully automated 1D Helmholtz coil setup that is internally water cooled, thus eliminating any confounding effects caused by temperature fluctuations. The coils also allow cells to be exposed to magnetic fields from multiple directions through automated controlled rotation. Preliminary data, acquired with the coils placed inside an incubator and on a rat vascular smooth muscle cell line, confirm previous reports that both microtubule and actin polymerization and dynamics are altered by weak magnetic fields.