Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
Rank 1
Structural bio & protein design
Biomolecular prediction shifts toward learned priors and coarse representations
Open-source models claim to exceed AlphaFold 3, while DEL-trained and coarse-prior models test generalization without full atomic input.
12 preprints · ↑ 3.5× vs 4-wk avg
Rank 2 Single-cell
Deep models infer tissue organization from spatial transcriptomics and imaging
Recent preprints extend spatial-omics deep learning to super-resolution, pathology-omics fusion, and cross-species immune foundation models.
14 preprints · 3 published · ↑ 3.3× vs 4-wk avg
Rank 3 Single-cell
Learned models forecast cell state transitions and perturbation responses
Models are moving from static prediction toward forecasting trajectories and benchmarked simulation of single-cell and regulatory behavior.
7 preprints · 1 published · ↑ 2.3× vs 4-wk avg
Rank 4 ML / AI methods
Machine learning predicts peptide, enzyme, and cyclization properties from sequence
Transfer learning, geometric deep learning, and standardized benchmarks are now applied to peptide, enzyme, and cyclization prediction.
2 preprints · 3 published · ↑ 2.1× vs 4-wk avg
Rank 5 Tools & infrastructure
Pipelines convert raw microbial measurements into interpretable biological calls
Recent preprints and a published pipeline address error correction and reproducibility in amplicon, metagenomic, metabolomic, and genomic microbial analyses.
4 preprints · 1 published · ↑ 2.0× vs 4-wk avg
Rank 6 Scalable EM and deep learning methods for genotype, splice, and methylation inference↑ 1.8×EpigenomicsRank 7 QTL and GWAS methods link noncoding variants to disease genes and targets↑ 1.7×GenomicsRank 8 Genome-wide sequence data inferring admixture, introgression, and population history↑ 1.8×GenomicsRank 9 Sequencing error and rate variation are modeled explicitly in genomic inference↑ 1.6×Tools & infrastructureRank 10 Sequence language models are judged on structural similarity and antigenic change outcomes↑ 1.6×ML / AI methodsRank 11 Learned models with explicit error control replace dataset-specific tuning in MS omics↑ 1.6×ProteomicsRank 12 Pretrained biological encoders are probed for interpretable, residue-level structure↑ 1.6×ML / AI methodsRank 13 Long-read and single-cell workflows resolve isoforms, splicing defects, and data artifacts↑ 1.6×Single-cellRank 14 Mechanistic evolution models used to infer protein structure and functional constraints↑ 1.8×Structural bio & protein designRank 15 Null-model scoring recovers multi-way co-localization and enhancer hub structure↑ 1.6×Single-cellRank 16 Natural-language and database layers for querying structured omics and biochemical data↑ 1.5×Tools & infrastructureRank 17 Embedding haplotype, chromatin and clonal structure into genomic inference modelsGenomicsRank 18 Supervised inference of pathogen resistance and identity from sequence and spectral dataML / AI methodsRank 19 New simulation and analysis software spans epidemic, neural and molecular modelsTools & infrastructureRank 20 Interaction rules and coupling structure shape tumor, metabolic, and developmental model…Systems biology
Show 19 moreShow less
Rank 21 ML on harmonized single-cell atlases nominates disease targets and cell statesSingle-cellRank 22 Comparative genomics links lineage-specific gene evolution to traits across animal taxaGenomicsRank 23 Compressed feature spaces are used to classify and curate cryo-EM particlesStructural bio & protein designRank 24 Supervised models infer latent biological state from molecular and clinical dataML / AI methodsRank 25 Mechanistic models map bistability, hysteresis and state transitions in cell regulationSystems biologyRank 26 Computational methods link noncoding and epigenetic variation to regulatory functionEpigenomicsRank 27 Statistical tools for power, differential expression and batch effects in omicsTools & infrastructureRank 28 Structure-constrained learned representations for medical and microscopy imagingML / AI methodsRank 29 AI and network models nominate delivery proteins and SARS-CoV-2 immune biomarkersSingle-cellRank 30 Statistical tools infer population size, motif enrichment and microbiome disease modelsTools & infrastructureRank 31 LLMs as reasoning and agentic components in drug discovery workflowsML / AI methodsRank 32 Restructured algorithms speed phylogenetic placement, RNA folding and graph alignmentTools & infrastructureRank 33 Graph and agent models predict perturbation, metabolic, and resistance biologyML / AI methodsRank 34 In silico tools map conjugated metabolites and standardize metabolomics annotationMetabolomicsRank 35 Functional and network models are assigning significance to uncertain genetic variantsGenomicsRank 36 Structural coarse-graining yields noise-robust fMRI connectivity from short scansSystems biologyRank 37 Simulations resolve membrane binding, RNA delivery, and enzymatic free-energy barriersStructural bio & protein designRank 38 Modeling how actin feedback and active-passive interfaces shape invasive cell behaviorSystems biologyRank 39 Browser-based platforms package AI molecular design and simulation workflowsTools & infrastructure