deep learning

Brain Decoding from a Deep Neural Network of Visual Memory Reveals the Maturation Trajectory of Memory Systems across Ages 3-7

The paper uses fMRI and a deep neural network to assess memory maturation in children aged 3-7, finding that hippocampal development drives memory pattern maturation.

NORMA 3.0: aesthetic and group-aware visualization and analysis of annotated biological networks

NORMA 3.0 provides enhanced visualization and analysis of annotated biological networks with improved group representation and layout algorithms.

An end-to-end in silico framework for truth-based validation and regression testing of clinical genome sequencing pipelines, from simulated reads to reportable findings

This paper introduces an end-to-end in silico framework to validate and test clinical genome sequencing pipelines from simulated reads to reportable findings using a fixed truth set.

Uncovering cross-signal interactions in single-cell and spatial omics

CRISP identifies and localizes interactions between biological signals in single-cell and spatial transcriptomic data, revealing their effects on tissue function and gene regulation.

A Distinct Mode of Chromosome-End Genome Instability Revealed by Circulome Characterization and Analyses

This paper identifies a new mode of chromosome-end instability in cancer using circular DNA analysis, revealing a distinct spatial signature called EDGE.

GOParsimony: parsimonious saltation across the Gene Ontology digraph for efficient functional annotation

GOParsimony filters GO terms to identify the most parsimonious functional annotations for gene/protein sets, outperforming existing methods in yeast and human data.

Sourmash accurately profiles strains from long-read metagenomes preserving gene content

This study evaluates Sourmash's ability to profile microbial strains from long-read metagenomes, showing accurate gene content recovery and abundance estimation.

Neural dynamics become significantly faster and more nonlinear at the peak of information processing

The paper shows that neural populations adjust their optimal temporal resolution during stimulus processing, becoming faster and more nonlinear when decoding information is most accurate.

Where single-cell foundation models help: a benchmark of annotation, perturbation and regulatory inference against simple baselines

This paper benchmarks single-cell foundation models against baselines for annotation, perturbation and regulatory inference using diverse datasets.

SeCAT (Sequence Consensus Amplicon Trimming): a simulation-informed Nextflow pipeline for cross-study harmonisation of 16S rRNA metabarcoding datasets

SeCAT is a Nextflow pipeline that harmonizes 16S rRNA metabarcoding datasets by identifying a consensus sequence window and testing its impact on community structure before merging datasets.

SPaRTA: Spatial Reasoning in Text-to-SQL for Omics Analysis

SPaRTA enables non-experts to query spatial-omics data via natural language, achieving 63-77% accuracy with fine-tuning on cancer tissue datasets.

MET inhibition exploits conserved regulatory dependencies in malignant rhabdoid tumors

This study identifies MET signaling as a targetable vulnerability in rhabdoid tumors through integrated single-cell multi-omic and spatial transcriptomics analyses.

Layer-Specific Source-Anchored Transfer Learning for Multi-Omics Survival Prediction in Younger Breast Cancer Patients

This paper introduces LSAT-EN, a transfer learning method that improves survival prediction for younger breast cancer patients using multi-omics data from older patients.

Photographic age gap is associated with frailty and predicts survival in laboratory mice

This paper uses photographs of mice to predict biological age and finds that older appearance correlates with frailty and shorter survival.

All-Atom Structure Tokenizer: AtomLM-9b for Biomolecular Structure Prediction

The paper introduces AtomLM-9b, a language model using an all-atom structure tokenizer to predict protein structures with performance comparable to AlphaFold 3.

Discovering compositional, interpretable structure in social behavior

This paper introduces a method to model social behavior using decomposed linear dynamical systems, showing it captures complex, compositional actions more effectively than existing methods.

Transmural Single-Nucleus Transcriptomics Nominates Candidate Disease Programs in Human Necrotizing Enterocolitis

The paper uses single-nucleus RNA sequencing to analyze intestinal tissues from NEC patients and controls, identifying disease-specific gene expression programs across multiple cell types.

Leveraging short and long-read sequencing to comprehensively characterize alternatively polyadenylated and alternatively spliced transcripts

The paper uses short and long-read sequencing to identify alternatively spliced and polyadenylated transcripts in mantle cell lymphoma, revealing novel splice variants and polyadenylation signals.

Explainable Multimodal Prognostic Modeling Uncovers New Clinically Relevant Pathological Features in a Multi-Institutional Pancreatic Neuroendocrine Tumor Cohort

This paper develops a multimodal prognostic model for pancreatic neuroendocrine tumors using clinicopathologic data and histology images, finding that stromal features are as prognostic as tumor cells.

Autonomous generalizability auditing of scientific foundation models

This paper introduces AutoSPECTRA, a protocol for auditing generalizability of scientific foundation models across biological domains.

XAMPGen: An Interpretable SI Pipeline For Novel Antimicrobial Peptide Discovery

This paper introduces XAMPGen, an interpretable generative model for discovering novel antimicrobial peptides using fine-tuned protein language models and explainable AI techniques.

2D U-NET Deep Learning Architecture for Atherosclerotic Lesion Segmentation and Quantification

This paper develops a 2D U-NET model to segment and quantify atherosclerotic lesions in mouse aortic sections, achieving high accuracy and significantly reducing measurement time.

Model-Agnostic Decomposition of the Difference in Mean Response Between Two Groups via Game-Theoretic Attribution with Minimum-Switching-Distortion Coupling

This paper introduces a model-agnostic decomposition method to analyze differences in mean response between groups, using game-theoretic attribution and optimal transport coupling.

Perturbation-induced gene programs map the hidden regulatory architecture of complex diseases

This paper introduces Perturbscape, a framework that uses perturbation-induced gene programs to identify disease-relevant genes and pathways, showing higher heritability informativeness than steady-state approaches.

GeneFamHunter: annotation-independent recovery of divergent and tandemly arrayed gene-family members from genome assemblies

GeneFamHunter recovers gene-family members from genome assemblies, outperforming existing methods in identifying divergent and tandemly arrayed genes.

Deep Recovery of Effective Migration Surfaces

This paper introduces dreems, a graph neural network that efficiently estimates effective migration surfaces from genetic data, achieving lower error and faster computation than existing methods.

Neural RNA splicing models rely on reading frame information unavailable to the splicing machinery

The paper shows that neural RNA splicing models rely on reading frame information that the splicing machinery cannot access, leading to errors in predicting poison exons.

RAPTORGraph: graph-based pathway modeling for causal discovery in single-cell perturbations

RAPTORGraph improves causal discovery in single-cell perturbations by modeling gene interactions and isolating perturbation effects, outperforming existing methods on downstream analyses.

Can the Human Plastisphere Account for the Rise of Type 1 Diabetes: A Computational Molecular Mimicry Study

This study investigates whether plastic fragments could induce type 1 diabetes through molecular mimicry, using AI and protein structure prediction.

Discovering Prognostic Risk Strata and Heterogeneous Progression Routes from Cross-Sectional Survival Data

GOTSurv models patient-specific disease progression from cross-sectional survival data, identifying heterogeneous risk strata and individualized progression trajectories without predefined risk cut-offs.