Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
ML / AI methods
Domain-informed molecular representations advance property and interaction prediction
Recent work encodes domain structure into molecular graphs and sequences for generation, reaction prediction and drug-target interaction tasks.
Single-cell
Generative and foundation models separate biological variation in single-cell data
Recent models encode explicit biological structure, such as circular cell-cycle subspaces and causal networks, directly into generative architectures.
Tools & infrastructure
Curated references and consensus statistics refine microbial profiling tools
Recent preprints and publications address reference database errors, limits of metagenomic functional recovery, and viral genome classification.
ML / AI methods
Label-efficient segmentation across histology, electron and multiplexed microscopy
Recent preprints apply self-supervised, self-training and foundation-model fine-tuning to nuclei, niche, ultrastructure and cell segmentation tasks.
ML / AI methods
Interpretable models predict regulatory activity and variant effects from DNA sequence
Genomic deep learning work increasingly pairs predictive accuracy with explanations of regulatory and variant-level outputs.