deep learning

Reinforcement Failing guides the discovery of emergent spatial dynamics in adaptive tumor therapy

This paper introduces Reinforcement Failing, an AI framework that identifies emergent spatial dynamics in adaptive tumor therapy by refining training environments.

A promptable foundation model enables automated multi-task dataset construction and real-time pose estimation in mice

The paper introduces DeepLabSAM for accurate pose estimation in interacting mice and constructs the MTMB dataset, showing improved performance over existing methods.

Atomistic Simulation of Blood Brain Barrier Permeability of Propolis Derived Natural Compounds

This study uses atomistic simulations to analyze BBB permeability of two propolis compounds, showing CAPE has favorable permeability while ARC faces significant resistance.

Automated detection, morphometric measurement, and size-class determination of harbor seals (Phoca vitulina) from UAV orthomosaics: a cascaded deep-learning pipeline with stratified validation across platforms, sensors, altitudes, and substrates

This paper develops a deep-learning pipeline for automated harbor seal detection and size-classification from drone imagery, showing high accuracy but variability in thermal vs. RGB performance.

Data mining sheds light on a novel family of plant-associated negative-sense RNA viruses linked to lispiviruses

The paper identifies a new family of plant-associated negative-sense RNA viruses using transcriptomic data and structural analysis, expanding known diversity by ~40-fold.

Comparative genomics and structure-informed analysis identify a putative tandem receptor-binding module in Pseudomonas phage UF_RH1

Phage phenotyping by measuring plaque expansion dynamics

This paper introduces a method to quantify phage killing rates by tracking plaque expansion dynamics on semi-solid media, revealing significant variation in expansion parameters across phages and taxonomic families.

Automated Segmentation and Tractography of Brainstem and Subcortical White Matter

The paper develops an automated pipeline for segmenting and parcellating brainstem and subcortical white matter pathways using diffusion MRI data.

VINTER: a generative vision-language model for annotating and interrogating spatial tissues

VINTER is a generative vision-language model that annotates spatial tissues by integrating histology, gene expression, and spatial neighbourhood, outperforming existing methods and transferring across datasets.

Application of deep learning and explainable AI-supported medical decision-making for facial phenotyping in genetic syndromes

This study evaluates if explainable AI (XAI) improves diagnostic accuracy and confidence in genetic syndrome recognition using facial images, finding mixed results with AI predictions having greater impact than XAI explanations.

tangermeme: A toolkit for understanding cis-regulatory logic using deep learning models

tangermeme provides tools to interpret deep learning models for cis-regulatory analysis, extracting human-readable insights from genomic predictions.

Hybrid Mechanistic-Neural Modeling of Concentration-Time Dynamics Generalizes Human Pharmacokinetics Prediction Across Unseen Chemical Space

PK-MUSE improves pharmacokinetics prediction by combining mechanistic models with neural corrections, achieving better performance across unseen chemical structures.

Cross-cohort Mamography]{Representation-dependent domain adaptation for cross-cohort generalization of mammography diagnostic models

This study evaluates cross-cohort generalization of mammography diagnostic models using different image representations and domain-adversarial learning, finding that MedSigLIP embeddings provide the best transferability.

Mutational fitness landscapes of diverse human influenza H1N1 neuraminidases

This study quantifies the replication fitness effects of mutations in diverse H1N1 influenza neuraminidases, revealing strain-dependent effects and epistasis patterns.

Nation-scale recovery of complete environmental plasmids pushes the boundaries of plasmid diversity through long-read sequencing

The paper expands plasmid diversity by analyzing long-read metagenomes, identifying novel plasmid lineages and providing tools for plasmid discovery.

Conserved synteny reveals a saccharolytic lifestyle across Kiritimatiellia

The paper uses conserved synteny analysis to reveal a shared saccharolytic lifestyle across Kiritimatiellia genomes.

CryoCodex: learning discrete structural representations for cryo-EM map post-processing

CryoCodex improves cryo-EM map quality by jointly enhancing density, estimating local quality, and masking molecules, outperforming existing methods on experimental maps.

Distinct Progression Paths Underlie the Diversity of Pancreatic Papillary Cystic Neoplasms

The paper identifies distinct progression paths in pancreatic papillary cystic neoplasms using integrated transcriptomics and histology, enabling phenotype-specific prediction of high-grade dysplasia.

How many crystal structures do you need to trust your docking results?

The paper evaluates how many crystal structures are needed for accurate pose prediction in drug discovery, finding that five structures per scaffold often suffice.

A lightweight deep-learning detector for the real-time monitoring of the invasive frogs Rhinella marina and Polypedates leucomystax

This paper develops a lightweight deep learning model for real-time detection of invasive frogs using acoustic data, achieving high accuracy on remote islands.

Desktop iDMS: A stand-alone application of intelligent differential mobility spectrometry (iDMS), a neural net-work that predicts optimal separation and compensa-tion voltages for field asymmetric differential ion mobil-ity spectrometry

The paper introduces iDMS, a neural network that predicts optimal separation and compensation voltages for DMS, improving glycosphingolipid analysis.

Deep-learning-based fMRI decoding of real-world size for hand-held objects

This study uses a 3D deep neural network to decode real-world object size from fMRI data, revealing a distributed neural network for subtle size differences beyond traditional visual areas.

A piece-wise linear recurrent neural network identifies generalizable dynamics from neural activity during decision-making in rodents

This study uses a piecewise linear recurrent neural network to identify generalizable latent dynamics in neural activity during decision-making in rodents, showing faster rotational dynamics for impulsive choices.

Generalizing Protein Fitness Prediction Across Distribution Shifts via Structural and Co-evolutionary Priors

This paper develops a framework for protein fitness prediction that improves generalization across distribution shifts by integrating structural and co-evolutionary information.

Single-Cell Behavioromics Enables Fate Prediction of Individual Hematopoietic Stem Cells

This study uses behavioromic analysis to predict hematopoietic stem cell fate with high accuracy and identifies Activin signaling's role in cell behavior and fate.

sORF-Trans2MS: a two-module deep learning framework for sORF translation and MS-supported microprotein prediction

sORF-Trans2MS is a two-module deep learning framework that predicts Ribo-supported sORFs and MS-supported microproteins, improving prediction and extending discovery to underrepresented candidates.

CelFDrive: Artificial Intelligence assisted microscopy for automated detection of rare events

CelFDrive automates high-resolution 3D imaging of cells using deep learning for efficient detection of rare events in large populations.

Powerful and accurate case-control analysis of spatial molecular data

This paper introduces VIMA, a method for analyzing spatial molecular data to identify disease-associated spatial features, showing improved performance over existing methods in simulations and real datasets.

Predicted structure of the complete CPLANE complex reveals novel interactors and mechanisms of Jbts17 during ciliogenesis

The paper predicts the structure of the CPLANE complex and identifies novel interactors of Jbts17, revealing its role in ciliogenesis through in silico and experimental validation.

Learning fly pose from vision-language annotations

This paper uses a vision-language model to generate annotations for fruit fly pose estimation, achieving high accuracy and enabling efficient training of pose estimation networks.