Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
and JavaScript.
A model-agnostic empirical framework is proposed to measure the uncertainty associated with protein embeddings and to assess the biological relevance of these embeddings in order to improve model reliability and performance on downstream tasks.
ROCKET improves experimental structure elucidation by integrating implicit structural knowledge from OpenFold, a trainable reimplementation of AlphaFold2, with X-ray crystallographic and cryo-electron microscopy experiments.
In this study, long-read RNA sequencing achieves accurate single-nucleotide polymorphism calling, haplotype phasing and allele-specific expression analysis.
To study myelination in vitro, a hydrogel-based micropillar array is used to mimic the stiffness and diameter of axons, with variations in these properties allowing evaluation of their influence. The platform can also be used to assess the effects of putative remyelination drugs.
A co-distillation framework is used to iteratively adapt sequence-only protein language models for high-accuracy variant effect prediction, without the need for additional structural or genetic data. Individual protein language models therefore self-improve by distilling the most confident predictions from multiple models, achieving state-of-the-art performance across multiple variant effect prediction benchmarks.
Hyperspectral fingerprint optoacoustic microscopy enables the differentiation of different lipid classes in living cells, which is of interest in the study of lipid metabolism in health and disease.
DIAMOND DeepClust provides an ultra-fast clustering method for organizing the protein universe of life at low sequence identity, enabling large-scale dimensionality reduction and improving downstream structure prediction with AlphaFold2.
An integrated mass spectrometry platform enabling automated collision-, electron- and photon-based fragmentation techniques is presented. A deep learning-based model trained to predict spectra across all dissociation methods further enhances protein identification.
CellVoyager is an artificial intelligence agent capable of exploring new biological hypotheses by autonomously analyzing single-cell RNA sequencing datasets and accounting for background information and prior analyses.
Connectome-seq allows access to connectomic and transcriptomic information in high throughput, using an RNA barcoding strategy and isolation of synaptosomes, followed by single-nucleus and single-synaptosome RNA sequencing.
ANNEVO advances accurate and scalable ab initio gene annotation of evolutionarily diverse genomes using deep learning approach modeling sequence evolution and long-range dependencies and mixture of experts (MoE) architecture.
SeeDB-Live is a tissue-clearing approach for live samples such as tissue slices or the in vivo brain. It improves image quality while having minimal effects on electrophysiological properties of neurons.
AF2BIND is a logistic regression model trained on AlphaFold2 pair features to predict small-molecule binding-site residues in proteins, without multiple sequence alignments, homology models or knowledge of the true ligand. AF2BIND was used to predict binding sites across the AF2-predicted human proteome, finding thousands of potentially new ligandable sites.
Engineered RNA domains that can adopt two- and fourfold symmetry are shown to be effective scaffolds for structure determination for RNAs otherwise too small for cryoEM.
Antscan is a publicly accessible database of synchrotron X-ray CT images of ants. The database covers almost 800 species from more than 200 genera and is coordinated with genome sequencing projects that will enable integrative analyses.
