Catálogo de publicaciones - revistas

<jats:p>Recent advances in machine learning have leveraged evolutionary information in multiple sequence alignments to predict protein structure. We demonstrate direct inference of full atomic-level protein structure from primary sequence using a large language model. As language models of protein sequences are scaled up to 15 billion parameters, an atomic-resolution picture of protein structure emerges in the learned representations. This results in an order-of-magnitude acceleration of high-resolution structure prediction, which enables large-scale structural characterization of metagenomic proteins. We apply this capability to construct the ESM Metagenomic Atlas by predicting structures for &gt;617 million metagenomic protein sequences, including &gt;225 million that are predicted with high confidence, which gives a view into the vast breadth and diversity of natural proteins.</jats:p>

<jats:p>Intercalated layered materials offer distinctive properties and serve as precursors for important two-dimensional (2D) materials. However, intercalation of non–van der Waals structures, which can expand the family of 2D materials, is difficult. We report a structural editing protocol for layered carbides (MAX phases) and their 2D derivatives (MXenes). Gap-opening and species-intercalating stages were respectively mediated by chemical scissors and intercalants, which created a large family of MAX phases with unconventional elements and structures, as well as MXenes with versatile terminals. The removal of terminals in MXenes with metal scissors and then the stitching of 2D carbide nanosheets with atom intercalation leads to the reconstruction of MAX phases and a family of metal-intercalated 2D carbides, both of which may drive advances in fields ranging from energy to printed electronics.</jats:p>

<jats:p>The initial evolution of warning signals in unprofitable prey, termed aposematism, is often seen as a paradox because any new conspicuous mutant would be easier to detect than its cryptic conspecifics and not readily recognized by naïve predators as defended. One possibility is that permanent aposematism first evolved through species using hidden warning signals, which are only exposed to would-be predators on encounter. Here, we present a large-scale analysis of evolutionary transitions in amphibian antipredation coloration and demonstrate that the evolutionary transition from camouflage to aposematism is rarely direct but tends to involve an intermediary stage, namely cryptic species that facultatively reveal conspicuous coloration. Accounting for this intermediate step can resolve the paradox and thereby advance our understanding of the evolution of aposematism.</jats:p>

<jats:p> It wasn’t that long ago when scientific collaboration between the United States and China was enthusiastically encouraged as a means to accomplish the best science. American universities established campuses in China, set up exchange programs for students and trainees, and hired highly productive Chinese researchers. That all changed in 2018, when then-President Trump launched the China Initiative to rid US academia of Chinese spies. As reporter Jeffrey Mervis describes in this issue of <jats:italic>Science</jats:italic> , the National Institutes of Health (NIH)—the largest federal funder of biomedical research—vigorously responded to this charge. The agency’s allegations and investigations have not only destroyed careers but also eroded trust in the agency and federal government across the scientific community. </jats:p>

<jats:p>Genetic sequences from Wuhan market may point to animal that spread SARS-CoV-2, but data remain hidden</jats:p>

<jats:p>Energy department opens new competition to run struggling Fermilab</jats:p>

<jats:p>Revamp aims to respond to U.S. challenge and promote “self-reliance” in science</jats:p>