Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
Rank 1
ML / AI methods
Encoding biological knowledge into compact sequence models for genomic and protein tasks
Alignment-free k-mer, taxonomy-guided, and structure-aware representations offer compact alternatives to alignment-based sequence analysis pipelines.
4 preprints · 3 published · ↑ 2.4× vs 4-wk avg
Rank 2 Genomics
Computational methods link variant effects to gene-level and trait-level inference
Variant-level evidence is being converted into gene-level and trait-level inference, with calibration and ancestry performance now assessed directly.
2 preprints · 3 published · ↑ 2.1× vs 4-wk avg
Rank 3 Single-cell
Computational methods infer spatial tissue organization from spatial transcriptomics data
Recent preprints extend spatial marker detection, domain discovery, slice alignment, and cell-trait mapping to imaging and spatial transcriptomics data.
8 preprints · 2 published · ↑ 2.1× vs 4-wk avg
Rank 4 Structural bio & protein design
Learned protein and peptide representations drive binding prediction and design
Recent preprints and papers test protein language model embeddings directly against binding, viability and design tasks.
5 preprints · 3 published · ↑ 2.0× vs 4-wk avg
Rank 5 Structural bio & protein design
Learned models predict biomolecular evolution, dynamics and variant effects
Zero-shot sequence models predict additive mutations well but fail under strong epistasis, constraining variant-effect prediction claims.
3 preprints · 3 published · ↑ 1.9× vs 4-wk avg
Rank 6 Long-read and hybrid inputs extend variant calling beyond short-read alignment↑ 1.8×GenomicsRank 7 Multimodal deep learning models validated against biological and clinical readouts↑ 1.7×ML / AI methodsRank 8 Structure-encoded inference of tumor states from single-cell, spatial, and plasma data↑ 1.6×Single-cellRank 9 Computational models infer regulatory and functional mechanisms from sequence and…↑ 1.5×ML / AI methodsRank 10 Ligand input format and charge alter co-folding predictions of protein-ligand complexes↑ 1.6×Structural bio & protein designRank 11 Experimental and reference data are coupling into protein structural modeling pipelines↑ 1.6×Structural bio & protein designRank 12 Multimodal embeddings and graph models score drug and target interactionsML / AI methodsRank 13 Statistical frameworks for peak, expression and methylation data across omics layersEpigenomicsRank 14 Genome-wide STRs and SNP-similarity methods refine population structure and phasingGenomicsRank 15 Cryo-EM and cryo-ET density constraining generative and regression-based structure…Structural bio & protein designRank 16 Intrinsic noise and protease competition shape oscillations and patterns in gene networksSystems biologyRank 17 Raw-read pipelines infer resistance determinants and pathogen hosts without assemblyGenomicsRank 18 Ancestry and environment shape immune eQTLs; simulators add HLA restrictionTools & infrastructureRank 19 Directed, scale-dependent and information-theoretic measures of brain network dynamicsSystems biologyRank 20 Benchmarking EEG, MEG and TMS Methods for Neural Machine LearningML / AI methods
Show 20 more
Rank 21 Software converts microscopy images into quantitative measurements across scalesTools & infrastructureRank 22 Benchmarks question whether single-cell foundation models capture regulatory mechanismSingle-cellRank 23 Generative and predictive models benchmarked for protein structure and dynamicsStructural bio & protein designRank 24 Simulations link mechanics of nucleosomes, capsids and membranes to functionStructural bio & protein designRank 25 Causal genetics and metabolic models nominate liver and aging disease targetsSystems biologyRank 26 New tools and benchmarks sharpen transcript-level RNA analysis across read typesTranscriptomicsRank 27 Chromosome-level genomes of three human Plasmodium species from natural infectionsGenomicsRank 28 Evaluation splits are being redesigned to expose shortcuts in biomolecular ML modelsML / AI methodsRank 29 Computational pipelines quantify single-molecule states in imaging and nanopore dataStructural bio & protein designRank 30 Explicit error models and learned components reshape MS proteomics data analysisProteomicsRank 31 Reusable software tools for alignment, regulatory-motif and knockout-phenotype analysisTools & infrastructureRank 32 Inferring pathogen resistance and variant abundance from indirect or mixed dataGenomicsRank 33 Model-based inference of host, gene, and community traits from metagenomic readsSystems biologyRank 34 Gene model curation and comparative annotation refine genomes across animals and algaeGenomicsRank 35 Selective compression and hashing cut the cost of long genomic sequence analysisTools & infrastructureRank 36 Agentic LLM pipelines build auditable, provenance-traced biomedical knowledge graphsTools & infrastructureRank 37 Faster exact and remote homology search tools for sequences and structuresTools & infrastructureRank 38 Post-mortem artifacts and cell-state complexity shape human transcriptomic profilesSingle-cellRank 39 Typing somatic mutations by regulatory, driver, repair and contextual impactGenomicsRank 40 Machine learning models link imaging, metabolite and network data to biological traitsML / AI methods