Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
Rank 1
Single-cell
Models infer single-cell state dynamics from lineage, regulatory, and perturbation data
Dynamical models now pair scRNA-seq trajectories with lineage context, regulatory structure, and biological constraints rather than static snapshots.
12 preprints · 3 published · ↑ 2.7× vs 4-wk avg
Rank 2 Genomics
Mapping non-coding and rare variants to gene regulation across tissues and cell types
Population-scale single-cell and multiomic resources now link variant effects to specific cell types, tissues and regulatory mechanisms.
7 preprints · 6 published · ↑ 2.6× vs 4-wk avg
Rank 3 Single-cell
Single-cell foundation models reused for perturbation and multi-omic prediction
Frozen single-cell foundation model representations are being reused via probes, distillation, and in-context learning for new prediction tasks.
12 preprints · 4 published · ↑ 2.4× vs 4-wk avg
Rank 4 Structural bio & protein design
Generative models capture interaction-induced conformational change in protein ensembles
Recent preprints extend conformational modeling from static bound geometries to unbound-to-bound transitions and experimentally weighted ensembles.
6 preprints · ↑ 2.3× vs 4-wk avg
Rank 5 Transcriptomics
Attributing gene expression to cell populations, cell states, and spatial niches
Spatial transcriptomics methods now model niche-conditioned gene co-expression and boundary-level label transitions rather than domain labels alone.
6 preprints · 2 published · ↑ 1.9× vs 4-wk avg
Rank 6 Curated resources and workflows unify spatial transcriptomics data across platforms↑ 1.9×Tools & infrastructureRank 7 Assay-aware and protocol-matched ML models for prioritizing small molecules↑ 1.9×ML / AI methodsRank 8 Multi-omics atlases and models link molecular profiles to clinical outcomes↑ 1.7×ML / AI methodsRank 9 Bounded agents and domain-specific software make biological workflows auditable↑ 1.7×Tools & infrastructureRank 10 Normalization, probe design and joint testing shape spatial and single-nucleus readouts↑ 1.6×Single-cellRank 11 Whole-genome doubling shapes tumour evolution and is detected at subclonal resolution↑ 1.6×GenomicsRank 12 Held-out validation design decides whether biological prediction gains survive scrutiny↑ 1.6×ML / AI methodsRank 13 Physical forces and diffusion models reproduce tissue architecture and branching in silicoSystems biologyRank 14 LLM agents bounded by source evidence and provenance in biomedical research↑ 1.6×ML / AI methodsRank 15 Pretrained biological models conditioned for biomolecular design and prediction tasksML / AI methodsRank 16 Deep learning methods now annotate chromatin architecture from Hi-C contact matricesEpigenomicsRank 17 Transcript-level methods refine isoform quantification from RNA-seq dataTranscriptomicsRank 18 Genome-scale data resolve organelle metabolism and sex chromosome evolutionGenomicsRank 19 Curated resources pair 3D biomolecular structures with experimental binding dataTools & infrastructureRank 20 Resources extend variant pathogenicity scoring to non-coding and mtDNA positionsTools & infrastructure
Show 20 more
Rank 21 Predicting spatial omics from histology and cross-modal data, judged by validation↑ 1.5×ML / AI methodsRank 22 Learned scores prioritize protein variants, binding pockets and complex interfacesStructural bio & protein designRank 23 Simulation and HDX-MS integration sharpen structural interpretation of protein dynamicsStructural bio & protein designRank 24 Statistical methods for microbiome inference under compositionality and confoundingML / AI methodsRank 25 Bursty production shapes timing, rhythm transmission and signaling informationSystems biologyRank 26 Deep learning models infer variants and taxa directly from raw sequencing dataGenomicsRank 27 Interpretable models map DNA sequence to initiation, co-accessibility and motifsEpigenomicsRank 28 Pangenome structure and transposon dynamics in genome expansion and contractionGenomicsRank 29 Simulations treat driven condensates as dynamic, feedback-controlled systemsStructural bio & protein designRank 30 LLM agents coupled to clinical data and genomic skills for outcome prediction and analysisML / AI methodsRank 31 Inferring glioblastoma invasion and proliferation parameters from imaging-based modelsSystems biologyRank 32 Predicted structures curated into reusable resources for proteome-scale annotationStructural bio & protein designRank 33 Curated knowledgebases organize microbial genomes by host or environmental contextGenomicsRank 34 Jagged-end correction for cfDNA methylation profiling preserves duplex integrityEpigenomicsRank 35 Integrating plasma metabolomics with genetic association and disease evidenceGenomicsRank 36 Computational frameworks for quantifying 3D and single-molecule microscopy dataTools & infrastructureRank 37 Computational prediction of ribozymes, communication residues, and DDX3X unwindingStructural bio & protein designRank 38 Parameter degeneracy and bursting transitions in minimal neuron and population modelsSystems biologyRank 39 Sequence models and assemblies map regulatory elements and selection footprintsGenomicsRank 40 Proteome-trained priors and benchmarks probe de novo peptide sequencing limitsProteomics