Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
Rank 1
ML / AI methods
Protein language model embeddings paired with structure and network context for…
Parameter-efficient adaptation of pretrained protein language models is now being paired with structure-aware evaluation across diverse prediction tasks.
12 preprints · 8 published · ↑ 3.2× vs 4-wk avg
Rank 2 Genomics
Long-read and single-cell methods quantify somatic mosaicism beyond nuclear SNVs
Long-read and single-cell data now allow per-locus quantification of somatic instability and mtDNA mosaicism, beyond conventional nuclear SNV calling.
8 preprints · 1 published · ↑ 3.2× vs 4-wk avg
Rank 3 Transcriptomics
Coupling histology and expression latent spaces for spatial transcriptomics analysis
Spatial transcriptomics tools increasingly pair expression with histology or protein data within shared latent models for tissue and cell-level prediction.
12 preprints · 1 published · ↑ 3.0× vs 4-wk avg
Rank 4 ML / AI methods
Transfer of learned structure across resolutions, domains, and biological priors
Few-shot and deconvolution-based transfer now links cell-line, bulk, and single-cell data to patient-level prediction tasks.
8 preprints · 3 published · ↑ 2.9× vs 4-wk avg
Rank 5 Single-cell
Statistical models of technical noise constrain inference across sequencing methods
Recent work applies Bayesian, mixture and variational models to dropout, copy-number and variant calling in single-cell and low-depth sequencing.
3 preprints · 5 published · ↑ 2.3× vs 4-wk avg
Rank 6 ML models prioritize drug candidates and targets under sparse ground truth↑ 2.3×Structural bio & protein designRank 7 Sequence-trained models target RNA structure, regulatory inference, and design↑ 2.2×ML / AI methodsRank 8 Bacterial defense and CRISPR records reveal mobile genetic element interactions↑ 2.1×GenomicsRank 9 Evolutionary timescales, spatial migration and trait architecture shape control outcomes↑ 1.9×Systems biologyRank 10 Generative models steer biosequence design toward multiple competing objectives↑ 1.6×Structural bio & protein designRank 11 Integrative modeling resolves flexible and multicomponent conformational ensembles↑ 1.5×Structural bio & protein designRank 12 Hardware-aware tools accelerate compression, error correction and variant analysis↑ 1.5×Tools & infrastructureRank 13 Local ancestry dosages and stability criteria refine genotype-based inference↑ 1.5×GenomicsRank 14 Computational tools and language models advance viral and phage genomic inferenceML / AI methodsRank 15 Transcriptomics and proteomics refine context-specific genome-scale metabolic models↑ 1.5×Systems biologyRank 16 Pangenomes and evolutionary models resolve structural and deleterious variation in cropsGenomicsRank 17 Spatial transcriptomics methods resolve tumor niches, immune subsets and epithelial statesTranscriptomicsRank 18 Molecular embeddings paired with biological and uncertainty componentsML / AI methodsRank 19 Joint latent models span brain connectomes, genomes, molecules, and single-cell statesML / AI methodsRank 20 Graph and integration software for single-cell and spatial transcriptomic analysisSingle-cell
Show 20 more
Rank 21 LLM agents autonomously run reproducible single-cell and multi-omics analysis pipelinesML / AI methodsRank 22 Learned latent spaces encode single-cell dynamics, perturbations, and fate decisionsSingle-cellRank 23 Latent and generative models reformulate protein design and MD sampling bottlenecksStructural bio & protein designRank 24 Transient fate-boundary programs and temporal history resolved in single-cell dataSingle-cellRank 25 Computational pipelines infer genes, ORFs and k-mer annotations from sequence dataGenomicsRank 26 Metapopulation models link mobility and surveillance sensitivity to infection dynamicsSystems biologyRank 27 Generative models design binders and antibodies from sequence or pocket geometryStructural bio & protein designRank 28 Alignment-free k-mer indexes extend classification, expression, and completeness scoringTools & infrastructureRank 29 Machine-learned and refined force fields for proteins, RNA, and peptidesStructural bio & protein designRank 30 Electrostatic and binding energy analyses inform membrane protein and receptor modelsStructural bio & protein designRank 31 Chromatin architecture and enhancer networks resolve context-dependent regulationEpigenomicsRank 32 Dynamical models of cellular processes span metabolism, circadian clocks and cell fateSystems biologyRank 33 Calibrated uncertainty for biological models under noisy, shifted, or dependent inputsML / AI methodsRank 34 Interpretable models link chromatin accessibility and enhancers to causal genesEpigenomicsRank 35 Ensemble models capture disordered protein and chromatin conformational heterogeneityEpigenomicsRank 36 Graph-level formalisms for pangenome diversity and homologyGenomicsRank 37 Structured biological knowledge graphs guide gene, protein and disease predictionSystems biologyRank 38 Simulations link molecular binding to emergent cell, matrix and tissue architectureStructural bio & protein designRank 39 Sub-3-bit subset rank and constant-space minimizers reduce k-mer index memory footprintTools & infrastructureRank 40 Cell-type eQTL mapping links non-coding brain variants to disease-relevant cellsSingle-cell