Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
Rank 1
Structural bio & protein design
Deep learning predicts biomolecular interactions, affinities, and interface quality
Pretrained protein and compound language models are now paired with graph and transformer architectures to extend learned predictors across interaction and…
7 preprints · 5 published · ↑ 2.3× vs 4-wk avg
Rank 2 ML / AI methods
Learned generative models map structured states under explicit scientific constraints
Recent preprints apply flow matching, diffusion, and Bayesian flow networks to antibody, single-cell, and molecular design problems.
4 preprints · 1 published · ↑ 2.1× vs 4-wk avg
Rank 3 Systems biology
Simulated transmission on reconstructed networks guides outbreak risk and surveillance
Simulations show 2 sequences per million people weekly would cut novel respiratory virus detection time by weeks to months.
1 preprint · 2 published · ↑ 2.0× vs 4-wk avg
Rank 4 Transcriptomics
Computational inference of spatial cell organization from spatial transcriptomics data
Benchmarks and new models now couple histology, spatial graphs, and denoising to resolve cell identity and tissue architecture from spatial data.
14 preprints · 5 published · ↑ 1.7× vs 4-wk avg
Rank 5 Epigenomics
Machine learning links tumor molecular profiles to outcome and epigenome editing
Predictive models now connect regulatory features such as RNA editing, silencers, and histone marks to clinical and expression outcomes.
2 preprints · 3 published · ↑ 1.7× vs 4-wk avg
Rank 6 Ionizable residue charge states gate conformational transitions in channels and complex I↑ 1.6×Structural bio & protein designRank 7 Ground-truth simulations and head-to-head comparisons benchmark microbiome tools↑ 1.5×Tools & infrastructureRank 8 Monoculture metabolic data and disordered models predict microbial community behavior↑ 1.5×Systems biologyRank 9 Learned models of MS2 spectra drive peptide identification and DIA quantification↑ 1.6×ProteomicsRank 10 Pangenome graphs and long-read methods expose structural variation in complex genomesGenomicsRank 11 Single-cell and multi-omics atlases map immune and tissue states in diseaseSingle-cellRank 12 Computational models turn brain and epigenomic data into individual risk estimatesEpigenomicsRank 13 Single-cell models infer regulatory dynamics and perturbation responses from noisy dataSingle-cellRank 14 Learned models predict antibody, nanobody, aptamer and enzyme properties from sequenceML / AI methodsRank 15 Computational protein-ligand binding from predicted and docked structuresStructural bio & protein designRank 16 Population genomic inference of selection and ancestry from partial or linked dataGenomicsRank 17 Sequence-to-function models decode regulatory grammar and 3D chromatin organizationEpigenomicsRank 18 Learning-based MRI methods address brain state, lesion, folding, and dementia tasksML / AI methodsRank 19 Foundation models and pooled screens map sequence-to-function design spaceML / AI methodsRank 20 ML and digital pathology enhance proteomic and cytometry data analysis workflowsProteomics
Show 16 more
Rank 21 Coarse-grained RNA models now capture sugar pucker, noncanonical pairing, and foldingStructural bio & protein designRank 22 Statistical models recover information lost to dropout and sparsity in expression dataGenomicsRank 23 Hi-C models reconstruct dynamic 3D genome folding; H2BK108Me2 may occlude a TSSEpigenomicsRank 24 Confinement and condensates reshape nucleic acid and protein folding in simulationStructural bio & protein designRank 25 Biological networks guide compound prioritization and drug repurposing predictionsSystems biologyRank 26 New tools place gene expression and spatial transcriptomic data in 3D anatomical contextTools & infrastructureRank 27 Phylogenetic tests show Asgard archaea dominate LECA eukaryotic gene originsGenomicsRank 28 Faster DNA language model tokenization, distance estimation, and sequence generationGenomicsRank 29 Generative models jointly optimize protein foldability, function, and enzyme kineticsStructural bio & protein designRank 30 Supervised models predict biomolecular features from sequence and variant dataML / AI methodsRank 31 Structure-derived and learned models sharpen remote homology and orthology assignmentTools & infrastructureRank 32 Fine-mapping and ancestry-diverse profiling narrow candidate causal variantsGenomicsRank 33 Sequence databases and search rules set the accuracy limits of proteomic identificationGenomicsRank 34 Language models and ontology priors are entering single-cell transcriptomic modelingSingle-cellRank 35 Chemical language models target peptide ADMET, nonstandard peptides, and metabolitesStructural bio & protein designRank 36 Microbial genomes reveal defense coevolution, cobamide ecology and antimicrobial peptidesGenomics