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.
DynaMight models continuous structural heterogeneity in cryo-EM datasets, leading to an improved reconstruction of the consensus structure. The study also explores the issue of overfitting when modeling structural flexibility.
This Perspective discusses the methodologies, application and evaluation of interpretable machine learning (IML) approaches in computational biology, with particular focus on common pitfalls when using IML and how to avoid them.
This Perspective discusses the issue of data leakage in machine learning based models and presents seven questions designed to identify and avoid the problems resulting from data leakage.
Smart parallel automated cryo-electron tomography (SPACEtomo) uses deep learning to fully automate data collection from lamella detection to tilt series acquisition, driving the future of cryo-ET through improved throughput and statistics.
Protein-tag degree of labeling (ProDOL) is a versatile reference-based approach for experimentally determining the degree of target labeling for improved protein counting and quantification and for optimizing labeling protocols in fixed and live cells.
Imaging the 4D choreography of subcellular events in living multicellular organisms at high spatiotemporal resolution could reveal life’s fundamental principles. Yet extracting these principles from petabyte-scale image data requires fusing advanced light microscopy and cutting-edge machine learning models with biological insight and expertise.
The dyes chosen for DNA-PAINT microscopy are pivotal for data quality. This Analysis shows a comprehensive comparison of 18 fluorescent dyes in DNA-PAINT and offers guidance for optimum dye selection in single-color and multiplexed imaging.
Rapid advancements in transcriptomics have enabled the quantification of individual transcripts for thousands of genes in millions of single cells. By coupling a machine learning inference framework with biophysical models describing the RNA life cycle, we can explore the dynamics driving RNA production, processing and degradation across cell types.
DeepPBS is a deep-learning model designed to predict the binding specificity of protein–DNA interactions using physicochemical and geometric contexts. DeepPBS functions across protein families and on experimentally determined as well as predicted protein–DNA complex structures.
CaST is a Ca2+-activated version of split-TurboID. The tool allows labeling active neurons quickly, simply by administration of exogenous biotin, thus enabling the study of behaviors that would be impaired by hardware required for the use of other, light-dependent tools.
Single-cell bisulfite sequencing enables the genome-wide quantification of DNA methylation at single-cell resolution, but methods to analyze the resulting data are lacking. The MethSCAn software accurately distinguishes cell types and states by scanning the genome for informative regions and providing a robust approach to quantifying methylation within these regions.
This work highlights the technical issues in previous approaches and introduces a preprocessing approach along with a software package, MethSCAn, for single-cell bisulfite sequencing data analysis.
We developed PINNACLE, a graph-based AI model for learning protein representations across cell-type contexts. These contextualized protein representations enable the integration of 3D protein structure with single-cell genomic-based representations to enhance protein–protein interaction prediction, analysis of drug effects across cell-type contexts, and prediction of therapeutic targets in a cell type-specific manner.
PINNACLE is a context-specific geometric deep learning model for generating protein representations. Leveraging single-cell transcriptomics combined with networks of protein–protein interactions, cell type-to-cell type interactions and a tissue hierarchy, PINNACLE generates high-resolution protein representations tailored to each cell type.
Bayesian nonparametric Track (BNP-Track) simultaneously determines emitter numbers and their tracks alongside uncertainty, extending the superresolution paradigm from static samples to single-particle tracking even in dense environments.