Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
Rank 1
ML / AI methods
Learned molecular and protein embeddings fused with graphs and language models
Recent preprints and JCIM papers pair pretrained protein and chemical language models with graph encoders and knowledge graphs for prediction tasks.
12 preprints · 4 published · ↑ 2.7× vs 4-wk avg
Rank 2 ML / AI methods
Deep learning links transcriptomic, histologic and single-cell data to biomedical outcomes
Recent preprints and publications apply deep models to pathway analysis, survival prediction, histology-to-expression mapping and single-cell integration.
9 preprints · 4 published · ↑ 2.3× vs 4-wk avg
Rank 3 ML / AI methods
Computational profiling of SARS-CoV-2 and influenza viral variation and host adaptation
Sequence-based and host-factor analyses now link SARS-CoV-2 and influenza variation to variant classification, antigenic prediction, and adaptation.
3 preprints · 2 published · ↑ 2.1× vs 4-wk avg
Rank 4 Genomics
Benchmarks show dataset-dependent accuracy for methods resolving human genomic variation
Recent preprints and publications benchmark variant, haplotype, and repeat methods against ground truth, exposing dataset-specific performance limits.
4 preprints · 4 published · ↑ 2.1× vs 4-wk avg
Rank 5 Tools & infrastructure
Modular, reproducible software pipelines and interactive tools for omics analysis
Omics analysis increasingly needs explicit, reproducible workflows, and recent tools pair that structure with LLM-assisted or interactive interfaces.
4 preprints · 4 published · ↑ 1.8× vs 4-wk avg
Rank 6 Objective-driven steering of generative models for RNA and peptide design↑ 1.8×Structural bio & protein designRank 7 Stochastic models infer selection and mutation history from single-cell and genomic data↑ 1.5×Single-cellRank 8 Interaction structure and external forcing govern multispecies dynamical states↑ 1.6×Systems biologyRank 9 Learned and geometric methods address noise and sparsity across omics data typesML / AI methodsRank 10 Cross-species single-cell and epigenomic profiling maps conserved regulatory programsSingle-cellRank 11 Deep learning and label-free imaging quantify cell state and tissue phenotypesML / AI methodsRank 12 Bounded geometric domains locate multi-omic signatures and tumor architecture statesSystems biologyRank 13 Learned models advance enzyme design, function annotation and binding-site predictionML / AI methodsRank 14 Reanalyzing existing data to adjudicate transcript structure and coding annotationTranscriptomicsRank 15 Learned and physics-informed methods resolve protein conformational states in cryo-EMStructural bio & protein designRank 16 Single-cell RNA-seq is now used to resolve tumor clonal structure and ecDNA selectionGenomicsRank 17 Deep and graph learning models screen glues, GPCR ligands, antibodies and neurotoxinsML / AI methodsRank 18 Deep learning models are individualizing brain maps from fMRI, MRI, and PETML / AI methodsRank 19 Kinetic, editing, and spatial structure enter transcriptomic models directlyTranscriptomicsRank 20 Image morphology and graph learning are used to infer molecular and spatial statesTranscriptomics
Show 20 more
Rank 21 Assembly-scale genomes link sex chromosome evolution, mobile elements and gene lossGenomicsRank 22 Annotation tools for repeats, isoforms, and polyadenylation sites are being benchmarkedGenomicsRank 23 New tools target metagenomic profiling, viral redundancy, and compositional biasTools & infrastructureRank 24 Computational methods compare protein structures beyond sequence-level similarityStructural bio & protein designRank 25 Explainable models rank methylation biomarkers; transcriptomes predict protein abundanceEpigenomicsRank 26 Genome drafts and orthogroup resources probe gene conservation and adaptationGenomicsRank 27 Curated knowledge bases unify biomarker, IBD, toxicity and drug-response data for…Tools & infrastructureRank 28 Metagenomic and metabolic models resolve microbiome functions in disease and environmentGenomicsRank 29 MD simulations link allosteric dynamics to ligand and antagonist binding modesStructural bio & protein designRank 30 Deep models for single-cell perturbation, annotation, and foundation-model interpretationSingle-cellRank 31 Network and neural methods map multi-target drug-target interactomesStructural bio & protein designRank 32 Analytical reliability of RNA-seq through QC, normalization, and assay benchmarkingTools & infrastructureRank 33 Structured host and population models sharpen epidemic thresholds and strain coexistenceSystems biologyRank 34 New methods target disorder, stickiness, low-complexity regions and dynamic interfacesStructural bio & protein designRank 35 Software frameworks link mass spectrometry imaging with spatial and multi-omics analysisTools & infrastructureRank 36 Hardware and algorithmic acceleration of sequence alignment and graph community detectionTools & infrastructureRank 37 Methods linking GWAS loci to biology via pleiotropy, epigenomes, GxE and epistasisGenomicsRank 38 Models of drug-tolerant persistence, cell-size control and leukaemia trial designSystems biologyRank 39 New mass spectrometry tools improve PTM quantitation and spectral library searchingProteomicsRank 40 Simulations link proton transport and hydration to pathogenic severityStructural bio & protein design