Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
Tools & infrastructure
Software and benchmarks for reproducible multi-omics analysis and interpretation
Large-scale benchmarks such as HADACA3 and scTranslation now test integration and translation methods across many pipelines and datasets.
Transcriptomics
Spatial neighborhood graphs drive imputation and multimodal spatial transcriptomics models
Neighborhood-based models now impute unmeasured genes, fill 3D slice gaps, and couple histology with transcriptomics across recent preprints and papers.
Structural bio & protein design
ML models advance protein design, stability, and interface prediction
Recent preprints benchmark pretrained and task-specific models against physics-based and baseline approaches across protein and DNA prediction tasks.
Single-cell
Learned models link single-cell dynamics, perturbations and spatial tissue structure
Foundation models, counterfactual editing and trajectory inference now target single-cell and spatial transcriptomics beyond annotation.
Transcriptomics
Biological priors built into graph and attention models for transcriptomic prediction
Recent preprints embed biological priors into KNN, Transformer, and graph models to make single-cell and perturbation predictions interpretable.