Computational models articles from across Nature Portfolio

Single-cell RNA-sequencing and spatial transcriptomics data enable the inference of how information is transmitted from one cell to another and how it modulates gene expression within cells. Now, a learning method infers networks describing how the inflow of one signal, mediated by intracellular gene modules, drives the outflow of other signals for intercellular communication.

According to a recent study, a small network consisting of four leaky integrate-and-fire neurons can reproduce the behavior of a single Hodgkin–Huxley neuron, thereby bridging the gap between endogenous and exogenous complexity.

scCM, a self-supervised contrastive learning method, effectively integrates large-scale CNS scRNA-seq data by clustering functionally related cells and separating dissimilar ones.

MMAPLE, a semi-supervised meta-learning framework, effectively addresses data scarcity and out-of-distribution challenges in deep learning. It uncovers novel drug-target, metabolite-enzyme, and microbiome-human interactions.

DNA-protein interactions form the fundamental structure of many pivotal biological processes. Here, authors propose ESM-DBP, which refines the DNA-binding protein sequence repertoire and domain-adaptive pretraining on general protein language models to predict DNA-binding protein-related tasks.

Direct electrophysiological recordings in the human prefronto-subthalamic circuit show that behavioral switching requires an optimal level of theta activity to be successful, which is well accounted for by a negative modulation of a drift-diffusion model starting point.

Existing 3D Spatial Transcriptomics reconstruction approaches often overlook spatial information or experiment-induced distortions. Here, authors propose ST-GEARS to bridge the gap between in vivo cell locations and in vitro analysis, accurately recovering spatial profiles.

Increasing the binding affinity of an antibody to its target antigen is key for antibody therapeutics. Here the authors report a pretrainable geometric graph neural network, GearBind, and explore its potential in in silico antibody affinity maturation.

Single-cell RNA-sequencing and spatial transcriptomics data enable the inference of how information is transmitted from one cell to another and how it modulates gene expression within cells. Now, a learning method infers networks describing how the inflow of one signal, mediated by intracellular gene modules, drives the outflow of other signals for intercellular communication.

According to a recent study, a small network consisting of four leaky integrate-and-fire neurons can reproduce the behavior of a single Hodgkin–Huxley neuron, thereby bridging the gap between endogenous and exogenous complexity.

By exploiting recent advances in modern artificial intelligence and large-scale functional genomic datasets, sequence-to-function models learn the relationship between genomic DNA and its multilayer gene regulatory functions. These models are poised to uncover mechanistic relationships across layers of cellular biology, which will transform our understanding of cis gene regulation and open new avenues for discovering disease mechanisms.

Methods for predicting bimolecular interactions are seeing tremendous growth, but challenges remain in capturing the full physical complexity of these interactions.

To address the challenge of pretraining foundational models with large datasets, a multi-task approach is proposed, thus helping to overcome the data scarcity problem in biomedical imaging.

A method leverages protein structural data to predict T-cell receptor–peptide interactions for unseen peptide epitopes, which can be particularly useful for applications in cancer immunotherapy, autoimmunity studies, and vaccine design.