Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
Rank 1
ML / AI methods
Sequence language models applied to RNA, protein, and enzyme prediction tasks
Pretrained RNA, protein, and nucleotide foundation models are being fused with auxiliary data and benchmarked on harder splits.
9 preprints · 3 published · ↑ 2.9× vs 4-wk avg
Rank 2 Epigenomics
Software frameworks couple chromatin architecture with transcriptomic and spatial profiles
Harmony2 scales single-cell integration to over 100 million cells, while recent tools add chromatin contacts to expression and accessibility analyses.
6 preprints · 6 published · ↑ 2.7× vs 4-wk avg
Rank 3 Genomics
Repeat-rich genomic contexts and pangenome variation inform eukaryotic genome evolution
Assemblies across thousands of genomes and 130 million plant transcripts now enable population-scale tests of genome architecture and function.
4 preprints · 3 published · ↑ 2.2× vs 4-wk avg
Rank 4 ML / AI methods
Pretrained protein language model embeddings paired with structural or graph context
Recent preprints and papers pair pretrained protein language model embeddings with structural or graph context for prediction tasks.
5 preprints · 6 published · ↑ 2.2× vs 4-wk avg
Rank 5 Genomics
Long-read sequencing quantifies somatic instability and mosaicism at tandem repeat loci
Long-read and Hi-C data now resolve haplotype-tagged somatic and mosaic states across repeats, cancer karyotypes and whole genomes.
4 preprints · 1 published · ↑ 2.1× vs 4-wk avg
Rank 6 Learned generative models replace costly sampling in protein dynamics and design↑ 2.1×Structural bio & protein designRank 7 Joint image-expression models refine spatial domains and single-cell representations↑ 1.8×Single-cellRank 8 Software advances in error correction, normalization, and throughput for sequencing data↑ 1.7×Tools & infrastructureRank 9 Learned sequence models co-optimize competing criteria in protein and DNA design↑ 1.6×Structural bio & protein designRank 10 Learned models rank drug candidates under limited labels and biological domain shift↑ 1.6×ML / AI methodsRank 11 Metagenomic records reveal bacterial interactions with mobile genetic elements in situ↑ 1.6×GenomicsRank 12 Machine-learned and physics-informed force fields refine biomolecular energeticsStructural bio & protein designRank 13 Gapped and hierarchical k-mer indexes drive alignment-free genomic analysis↑ 1.6×Tools & infrastructureRank 14 Interpretable models rank causal genes from GWAS, QTL, and genomics dataSystems biologyRank 15 Automated computational pipelines for structure-based design inputs and scoringStructural bio & protein designRank 16 Latent embeddings fuse sequence, structure and graph views for biomolecular predictionML / AI methodsRank 17 Language models and probabilistic inference target viral and phage genome analysisML / AI methodsRank 18 Benchmarking genotype-to-phenotype inference methods for rigor and implementationML / AI methodsRank 19 Evolutionary timescales and heritable traits alter epidemic and tumor model outcomesSystems biologyRank 20 Single-nucleus eQTL and enhancer analyses link brain variants to cell-type expressionGenomics
Show 20 more
Rank 21 Multimodal models link histology and spatial transcriptomics to tumor statesML / AI methodsRank 22 Electrostatics and binding energetics test membrane protein and peptide model predictionsStructural bio & protein designRank 23 Computational prediction of drug off-target binding and phosphoproteomic signaling…ProteomicsRank 24 Long-read transcript assemblies and isoform-level annotation refine transcript-resolved…TranscriptomicsRank 25 Chromatin context and transcription factor networks define regulatory programsEpigenomicsRank 26 Optimal transport and neural mass models map stimulus-driven brain routing and controlSystems biologyRank 27 Coarse-grained models reveal ion, confinement, and search effects on DNA and PARStructural bio & protein designRank 28 Language-model agents execute reproducible omics and biological knowledge-graph queriesTools & infrastructureRank 29 Open-source pipelines turn spatial omics and whole-slide images into reusable workflowsTools & infrastructureRank 30 Ensemble models recapitulate chromatin packing and Micro-C contact landscapesEpigenomicsRank 31 Ensemble modeling against experimental restraints resolves flexible protein complexesStructural bio & protein designRank 32 Generative transformers and flow models predict single-cell responses to perturbationSingle-cellRank 33 Language models are generating mRNA-level antibody and coding sequencesML / AI methodsRank 34 Structure-based energetics and multiscale simulation predict membrane receptor behaviorStructural bio & protein designRank 35 Space-time tradeoffs and minimizer designs cut memory for k-mer indexing and filteringTools & infrastructureRank 36 New tools quantify transposable elements, structural variants and RNA modificationsGenomicsRank 37 Latent models infer cellular dynamics from single-cell and spatial snapshot dataSingle-cellRank 38 Curated atlases and compendia unify cell, ligase, glycan and trait-variant dataTools & infrastructureRank 39 Format-aware compression and SIMD parsing target genomic and structure file handlingTools & infrastructureRank 40 GxE-aware genomic prediction improves Miscanthus biomass yield forecastsGenomics