Catálogo de publicaciones - revistas

<jats:title>Abstract</jats:title><jats:p>The magnetic dynamo cycle of the Sun features a distinct pattern: a propagating region of sunspot emergence appears around 30° latitude and vanishes near the equator every 11 years (ref. <jats:sup>1</jats:sup>). Moreover, longitudinal flows called torsional oscillations closely shadow sunspot migration, undoubtedly sharing a common cause<jats:sup>2</jats:sup>. Contrary to theories suggesting deep origins of these phenomena, helioseismology pinpoints low-latitude torsional oscillations to the outer 5–10% of the Sun, the near-surface shear layer<jats:sup>3,4</jats:sup>. Within this zone, inwardly increasing differential rotation coupled with a poloidal magnetic field strongly implicates the magneto-rotational instability<jats:sup>5,6</jats:sup>, prominent in accretion-disk theory and observed in laboratory experiments<jats:sup>7</jats:sup>. Together, these two facts prompt the general question: whether the solar dynamo is possibly a near-surface instability. Here we report strong affirmative evidence in stark contrast to traditional models<jats:sup>8</jats:sup> focusing on the deeper tachocline. Simple analytic estimates show that the near-surface magneto-rotational instability better explains the spatiotemporal scales of the torsional oscillations and inferred subsurface magnetic field amplitudes<jats:sup>9</jats:sup>. State-of-the-art numerical simulations corroborate these estimates and reproduce hemispherical magnetic current helicity laws<jats:sup>10</jats:sup>. The dynamo resulting from a well-understood near-surface phenomenon improves prospects for accurate predictions of full magnetic cycles and space weather, affecting the electromagnetic infrastructure of Earth.</jats:p>

<jats:title>Abstract</jats:title><jats:p>The development of neural techniques has opened up new avenues for research in machine translation. Today, neural machine translation (NMT) systems can leverage highly multilingual capacities and even perform zero-shot translation, delivering promising results in terms of language coverage and quality. However, scaling quality NMT requires large volumes of parallel bilingual data, which are not equally available for the 7,000+ languages in the world<jats:sup>1</jats:sup>. Focusing on improving the translation qualities of a relatively small group of high-resource languages comes at the expense of directing research attention to low-resource languages, exacerbating digital inequities in the long run. To break this pattern, here we introduce No Language Left Behind—a single massively multilingual model that leverages transfer learning across languages. We developed a conditional computational model based on the Sparsely Gated Mixture of Experts architecture<jats:sup>2–7</jats:sup>, which we trained on data obtained with new mining techniques tailored for low-resource languages. Furthermore, we devised multiple architectural and training improvements to counteract overfitting while training on thousands of tasks. We evaluated the performance of our model over 40,000 translation directions using tools created specifically for this purpose—an automatic benchmark (FLORES-200), a human evaluation metric (XSTS) and a toxicity detector that covers every language in our model. Compared with the previous state-of-the-art models, our model achieves an average of 44% improvement in translation quality as measured by BLEU. By demonstrating how to scale NMT to 200 languages and making all contributions in this effort freely available for non-commercial use, our work lays important groundwork for the development of a universal translation system.</jats:p>

<jats:title>Abstract</jats:title><jats:p>Histone acetylation regulates gene expression, cell function and cell fate<jats:sup>1</jats:sup>. Here we study the pattern of histone acetylation in the epithelial tissue of the <jats:italic>Drosophila</jats:italic> wing disc. H3K18ac, H4K8ac and total lysine acetylation are increased in the outer rim of the disc. This acetylation pattern is controlled by nuclear position, whereby nuclei continuously move from apical to basal locations within the epithelium and exhibit high levels of H3K18ac when they are in proximity to the tissue surface. These surface nuclei have increased levels of acetyl-CoA synthase, which generates the acetyl-CoA for histone acetylation. The carbon source for histone acetylation in the rim is fatty acid β-oxidation, which is also increased in the rim. Inhibition of fatty acid β-oxidation causes H3K18ac levels to decrease in the genomic proximity of genes involved in disc development. In summary, there is a physical mark of the outer rim of the wing and other imaginal epithelia in <jats:italic>Drosophila</jats:italic> that affects gene expression.</jats:p>

<jats:title>Abstract</jats:title><jats:p>All drugs of abuse induce long-lasting changes in synaptic transmission and neural circuit function that underlie substance-use disorders<jats:sup>1,2</jats:sup>. Another recently appreciated mechanism of neural circuit plasticity is mediated through activity-regulated changes in myelin that can tune circuit function and influence cognitive behaviour<jats:sup>3–7</jats:sup>. Here we explore the role of myelin plasticity in dopaminergic circuitry and reward learning. We demonstrate that dopaminergic neuronal activity-regulated myelin plasticity is a key modulator of dopaminergic circuit function and opioid reward. Oligodendroglial lineage cells respond to dopaminergic neuronal activity evoked by optogenetic stimulation of dopaminergic neurons, optogenetic inhibition of GABAergic neurons, or administration of morphine. These oligodendroglial changes are evident selectively within the ventral tegmental area but not along the axonal projections in the medial forebrain bundle nor within the target nucleus accumbens. Genetic blockade of oligodendrogenesis dampens dopamine release dynamics in nucleus accumbens and impairs behavioural conditioning to morphine. Taken together, these findings underscore a critical role for oligodendrogenesis in reward learning and identify dopaminergic neuronal activity-regulated myelin plasticity as an important circuit modification that is required for opioid reward.</jats:p>

<jats:title>Abstract</jats:title><jats:p>The capacity for terrestrial ecosystems to sequester additional carbon (C) with rising CO<jats:sub>2</jats:sub> concentrations depends on soil nutrient availability<jats:sup>1,2</jats:sup>. Previous evidence suggested that mature forests growing on phosphorus (P)-deprived soils had limited capacity to sequester extra biomass under elevated CO<jats:sub>2</jats:sub> (refs. <jats:sup>3–6</jats:sup>), but uncertainty about ecosystem P cycling and its CO<jats:sub>2</jats:sub> response represents a crucial bottleneck for mechanistic prediction of the land C sink under climate change<jats:sup>7</jats:sup>. Here, by compiling the first comprehensive P budget for a P-limited mature forest exposed to elevated CO<jats:sub>2</jats:sub>, we show a high likelihood that P captured by soil microorganisms constrains ecosystem P recycling and availability for plant uptake. Trees used P efficiently, but microbial pre-emption of mineralized soil P seemed to limit the capacity of trees for increased P uptake and assimilation under elevated CO<jats:sub>2</jats:sub> and, therefore, their capacity to sequester extra C. Plant strategies to stimulate microbial P cycling and plant P uptake, such as increasing rhizosphere C release to soil, will probably be necessary for P-limited forests to increase C capture into new biomass. Our results identify the key mechanisms by which P availability limits CO<jats:sub>2</jats:sub> fertilization of tree growth and will guide the development of Earth system models to predict future long-term C storage.</jats:p>

<jats:title>Abstract</jats:title><jats:p>Increasing rates of autoimmune and inflammatory disease present a burgeoning threat to human health<jats:sup>1</jats:sup>. This is compounded by the limited efficacy of available treatments<jats:sup>1</jats:sup> and high failure rates during drug development<jats:sup>2</jats:sup>, highlighting an urgent need to better understand disease mechanisms. Here we show how functional genomics could address this challenge. By investigating an intergenic haplotype on chr21q22—which has been independently linked to inflammatory bowel disease, ankylosing spondylitis, primary sclerosing cholangitis and Takayasu’s arteritis<jats:sup>3–6</jats:sup>—we identify that the causal gene, <jats:italic>ETS2</jats:italic>, is a central regulator of human inflammatory macrophages and delineate the shared disease mechanism that amplifies <jats:italic>ETS2</jats:italic> expression. Genes regulated by ETS2 were prominently expressed in diseased tissues and more enriched for inflammatory bowel disease GWAS hits than most previously described pathways. Overexpressing <jats:italic>ETS2</jats:italic> in resting macrophages reproduced the inflammatory state observed in chr21q22-associated diseases, with upregulation of multiple drug targets, including TNF and IL-23. Using a database of cellular signatures<jats:sup>7</jats:sup>, we identified drugs that might modulate this pathway and validated the potent anti-inflammatory activity of one class of small molecules in vitro and ex vivo. Together, this illustrates the power of functional genomics, applied directly in primary human cells, to identify immune-mediated disease mechanisms and potential therapeutic opportunities.</jats:p>