contrastive learning

HERACLES for fast and accurate likelihood-based reconstruction of massive lineages

HERACLES improves lineage reconstruction accuracy and scalability by optimizing cell coordinates in hyperbolic space using a composite likelihood model of CRISPR editing.

GeneEmbedLLM: Harnessing Large Language Models for Gene Functional Embedding and Association Prediction

This paper introduces GeneEmbedLLM, a framework using large language models to generate gene embeddings from multi-source textual annotations, achieving competitive performance in various downstream tasks.

Contrastive Learning Improves Generalization on M2OR Olfactory Binding Dataset

The paper evaluates contrastive learning on VOC-OR binding datasets, showing improved generalization and reduced performance inflation from identity leakage.

Language-guided domain adaptation improves protein sequence rep-resentations for antimicrobial resistance analysis

BacLLM improves protein sequence representations for antimicrobial resistance analysis by integrating functional knowledge and enhancing resistance gene identification and mechanism prediction.

Reference-Guided De Novo Retrieval of 3D Chromatin Organizational Features with Representation Learning

The paper introduces C-Fetch, a framework that retrieves 3D chromatin features using reference examples, showing improved enrichment for biologically relevant contacts compared to existing methods.

Error-Controlled Discovery of Protein Interactions via Conformal Calibration

PPI-Caliper controls false discovery rate in protein interaction prediction using conformal calibration without retraining models.

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.

RELAY: Relational In-Context Learning for Cellular Perturbation Prediction

RELAY predicts cellular responses to perturbations by learning relational patterns from measured data, outperforming existing methods on unseen variable values and combinations.

FedEpitope: Parameter-Efficient Federated Protein Language Models for Linear B-Cell Epitope Prediction

FedEpitope uses federated learning with LoRA to predict linear B-cell epitopes without centralizing data, achieving good performance across multiple clients.

MaxCAT-Net: Hybrid Transformer-Based Skin Lesion Classification and Lightweight Segmentation with Calibration and Cross-Dataset Validation

MaxCAT-Net improves skin lesion classification and segmentation with hybrid architecture and calibration, showing performance on benchmark datasets but with preliminary results.

A single-cell multimodal foundation model enables cross modality analysis between transcriptome and chromatin

MUTUAL enables cross-modality analysis between transcriptome and chromatin by learning a shared RNA-ATAC space from multiome data without refitting.

Ground-truth evaluation of sparse autoencoder features on spatial, single-cell and language-model embeddings

The paper evaluates sparse autoencoder features on various embeddings, finding that reproducibility does not guarantee meaningfulness, and proposes reporting features with concept coverage and purity.

Auditory tracking magnitude, temporal alignment, and anatomical distribution during wakefulness and propofol-based anesthesia

The paper examines how auditory tracking changes during wakefulness and propofol anesthesia by analyzing response magnitude, temporal alignment, and anatomical distribution using sEEG data.

Neopep: Reinforcement Learning From Machine-Learned Force Field Energy for Non-Canonical Peptide Structure Prediction

Neopep uses reinforcement learning with machine-learned force field energy to predict structures of non-canonical amino acid-containing peptides, achieving lower RMSD than existing methods.

Regularized logistic regression-based gene selection for functional analysis of heterogeneous scRNA-seq macrophage populations in muscle aging and regeneration.

The paper introduces DisTwiner, a method for subgroup-aware gene selection in scRNA-seq data, showing it captures biologically relevant aging pathways in macrophages.

VenusGenome: Genome-Aware Masked Autoencoding for Microbial Genome Representation Learning

VenusGenome improves microbial genome representation learning by using a genome-aware masked autoencoder, achieving high performance in phenotype prediction and gene-level tasks.

How Much Can AI Virtual Cells Gain from Unified Perturbational Modeling?

This paper evaluates how unified perturbational modeling improves AI virtual cell predictors by integrating heterogeneous Perturb-seq datasets and achieving better performance across benchmarks.

Prediction transfer does not imply reliability transfer in single-cell foundation models

The paper investigates reliability transfer in single-cell foundation models across biological contexts and finds it varies independently from prediction transfer.

Fine-Grained and Reliable Cell Classification from Pathology Images

CellBridge improves cell classification in pathology images by transferring molecular information, achieving higher accuracy on fine-grained labels.

LoCoFlow: Local-Context Flow Matching for Single-Cell Dynamics and Fate Inference

LoCoFlow improves trajectory inference in single-cell data by incorporating local geometry and fate uncertainty, achieving better performance than existing methods.

Systematic comparison of color representations between humans and deep neural networks: towards predicting human color perception in a vast color space

This study compares how different learning paradigms enable deep neural networks to predict human color perception across a vast color space.

Regional history outweighs phylogeny in shaping recent demographic trajectories of Indo-West Pacific sea snakes

This study shows that regional history, not phylogeny or range size, primarily shapes demographic changes in Indo-West Pacific sea snakes over the last 200 thousand years.

DNA-LLM2Vec: Architecture-Aware Task-Conditioned Adaptation of Decoder DNA Foundation Models

DNA-LLM2Vec adapts decoder DNA models to bidirectional context for genomic tasks, improving performance on extended-context tasks but varying by task and backbone.

Revealing pan-cancer clonal niches of T cells from single-cell RNA sequencing using contrastive learning

CloneTrast uses contrastive learning to infer T-cell clonal niches from single-cell RNA-seq, identifying NK-like niches linked to poor immunotherapy response across cancers.

LOCALE: Local-Alignment Embeddings for Noise-Robust DNA Search at Scale

LOCALE improves DNA sequence search by learning variation-robust embeddings, outperforming existing methods on benchmark datasets and real biological variance tasks.

REMAG: recovery of eukaryotic genomes from metagenomic data using non-contrastive self-supervised learning

REMAG improves eukaryotic genome recovery from metagenomic data by using non-contrastive self-supervised learning, outperforming existing tools in producing near-complete genomes.

An integrated multi-omic atlas of the choroid plexus highlights conserved molecular programs of epithelial function

This study integrates multi-omic data from mouse and human choroid plexus to identify conserved molecular programs of epithelial function across species.

Histology-aware Inference of Cell-cell Communication in Spatial Transcriptomics

This paper introduces HARMONIC, a framework that improves cell-cell communication inference in spatial transcriptomics by incorporating histology data, showing it better captures anatomical context.

From Sparse Visits to Continuous Molecular Trajectories with Flow Matching

CohortFM reconstructs continuous molecular trajectories from sparse longitudinal data, revealing biologically meaningful progression patterns across diseases.

CRISCross: Integration of Epigenetic Information for Deep-Learning-Based Prediction of CRISPR-Cas9 Off-Targets Across Cell Types

CRISCross improves CRISPR-Cas9 off-target prediction by integrating epigenetic features with sequence data across cell types, showing better performance than sequence-only models.