contrastive learning

GatorST: A Versatile Contrastive Meta-Learning Framework for Spatial Transcriptomic Data Analysis

GatorST improves spatial transcriptomic analysis by integrating graph-based modeling, pseudo-labeling, and contrastive meta-learning to enhance spatial domain identification and downstream tasks.

NEAR: Neural Embeddings for Amino acid Relationships

NEAR improves homology detection speed and accuracy by using neural embeddings trained with contrastive learning, outperforming existing methods as a pre-filter for protein sequence annotation.

Structural Repetition Detector: multi-scale quantitative mapping of molecular complexes through microscopy

The paper introduces SReD, an unsupervised framework for detecting repetitive biological structures in microscopy images, demonstrating its use on three datasets.

PathCLAST: Pathway-Augmented Contrastive Learning with Attention for Spatial Transcriptomics

PathCLAST improves spatial domain identification in spatial transcriptomics by integrating gene expression, histopathology, and pathway information, achieving better clustering accuracy than existing methods.

CryoDRGN-AI: Neural ab initio reconstruction of challenging cryo-EM and cryo-ET datasets

The paper introduces cryoDRGN-AI, a method for ab initio 3D reconstruction of cryo-EM and cryo-ET datasets, revealing new conformational states and unresolved motions.

Unsupervised Multi-scale Segmentation of Cellular Cryo-electron Tomograms with Stable Diffusion Foundation Model

The paper introduces an unsupervised method for segmenting subcellular structures in cryo-ET images using a Stable Diffusion model, achieving performance close to expert annotations.

A DNA Sequence Imaging Approach to Predict Splice Sites Using Deep Learning

This paper introduces a DNA sequence imaging method for splice site prediction, showing that DFCP outperforms FCGR using deep learning on Arabidopsis and human data.

The Topology of Representational Geometry

This paper introduces a method that enhances representational similarity analysis by incorporating computational topology to better capture the shape of neural representational spaces.

MFPLI: A Computational Framework for Assessing Biological Authenticity of Protein-Ligand Interactions Using Molecular Fingerprints and Structural Features

MFPLI evaluates protein-ligand binding authenticity using deep learning with physicochemical and structural features, achieving high accuracy in distinguishing native conformations from docking poses.

A conserved immune dysregulation is associated with infection severity, risk factors prior to infection, and treatment response

The paper identifies a conserved immune signature linked to infection severity, risk factors, and treatment response using multi-cohort transcriptomic and proteomic data.

Predicting Spatial Transcriptomics from H&E Image by Pretrained Contrastive Alignment Learning

This paper introduces CarHE, a method that predicts spatial transcriptomics data from H&E images using contrastive learning, achieving high accuracy and enabling cost-effective analysis.

PB-scope: Contrastive learning of dynamic processing body formation reveals undefined mechanisms of approved compounds

PB-scope uses deep learning to analyze PB dynamics and identifies drugs affecting JAK signaling pathway.

NiCLIP: Neuroimaging contrastive language-image pretraining model for predicting text from brain activation images

NiCLIP predicts cognitive tasks and domains from brain activation maps using a contrastive language-image model trained on neuroscientific articles.

Rethinking Representation Complexity in Drug-Target Prediction via Supervised Vector Quantization

This study evaluates complex features in drug-target interaction prediction and introduces a supervised vector quantization framework that improves performance and interpretability.

Squidly: Enzyme Catalytic Residue Prediction Harnessing a Biology-Informed Contrastive Learning Framework

Squidly predicts catalytic residues using a biology-informed contrastive learning framework, outperforming existing methods and integrating with BLAST for efficient annotation.

Segment AnyNeuron

Segment AnyNeuron is an active learning framework for neuron segmentation using minimal annotations, achieving high precision on diverse neuron imaging datasets.

Applying Self-Supervised Tissue Patch Encoders to Enable Cellular Graph Encoding

This study shows that a pretrained tissue patch encoder can effectively encode cellular graphs, capturing cell-level features and spatial relationships in tissue patches.

CSLAN: Cross-Species Latent Alignment Network for Trauma-Related Cell-Type Classification

CSLAN transfers mouse cell type knowledge to human trauma-related scRNA-seq data, achieving 96.67% accuracy in immune cell classification.

Simpatico: accurate and ultra-fast virtual drug screening with atomic embeddings

Simpatico improves virtual drug screening by using atomic-level embeddings to rapidly and accurately identify potential ligands for protein targets.

Pisces: A multi-modal data augmentation approach for drug combination synergy prediction

Pisces enhances drug combination synergy prediction by generating 64 augmented views per drug pair, achieving state-of-the-art results across multiple prediction tasks.

HEAL-KGGen: A Hierarchical Multi-Agent LLM Framework with Knowledge Graph Enhancement for Genetic Biomarker-Based Medical Diagnosis

HEAL-KGGen improves LLM performance in genetic diagnostics by integrating a knowledge graph with a multi-agent framework for multi-omics data interpretation.

Molecular Subtyping Based on Hippocampal Cryptic Exon Burden Reveals Proteome-wide Changes Associated with TDP-43 Pathology across the Spectrum of LATE and Alzheimer's Disease

This study identifies molecular subtypes based on cryptic exon burden in hippocampal tissue, revealing proteomic changes linked to TDP-43 pathology in LATE and Alzheimer's disease.

New idtracker.ai: rethinking multi-animal tracking as a representation learning problem to increase accuracy and reduce tracking times

The paper introduces a new idtracker.ai that improves multi-animal tracking by reframing it as a representation learning problem, enhancing accuracy and reducing tracking time.

RETINA: Reconstruction-based Pre-Trained Enhanced TransUNet for Electron Microscopy Segmentation on the CEM500K Dataset

RETINA improves EM image segmentation by combining pre-training on CEM500K with a hybrid model, outperforming existing methods on five public datasets.

Error reduction in leukemia machine learning classification with conformal prediction

This study applies conformal prediction to reduce false negative rates in leukemia subtype classification using RNA-seq data.

Contrastive functional connectivity defines neurophysiology-informed symptom dimensions in major depression

This paper identifies two neurophysiology-informed symptom dimensions in MDD using contrastive functional connectivity analysis of fMRI data.

A Self-Supervised Learning Approach for High Throughput and High Content Cell Segmentation

This paper introduces a self-supervised learning method for cell segmentation that outperforms Cellpose in accuracy and reduces reliance on labeled data and computational expertise.

Robust Deep Denoising of Soft X-Ray Tomography Data for Biological Research: Targeting Tilt Series Versus Reconstructed Tomograms

This study compares denoising raw tilt series versus reconstructed tomograms in soft X-ray tomography, finding that denoising tilt series provides more reliable results and faster inference.

DPAC: Prediction and Design of Protein-DNA Interactions via Sequence-Based Contrastive Learning

DPAC predicts and designs protein-DNA interactions using sequence-based contrastive learning, outperforming structure-based methods in speed and accuracy.

AI4CellFate: Interpretable Early Cell Fate Prediction with Generative AI

AI4CellFate predicts cell fate from microscopy timelapses using generative AI and contrastive learning, enabling early prediction with limited annotation.