Top 40
View all ›BACKFILLED Code and industry signals are partial for backfilled days.
Rank 1
Transcriptomics
Computational methods impute, deconvolve and align spatial omics data across modalities
Spatial omics datasets are sparse, multimodal and increasingly 3D, driving new imputation, uncertainty and integration methods.
12 preprints · 2 published · ↑ 2.4× vs 4-wk avg
Rank 2 ML / AI methods
Sparse autoencoders expose biological features in single-cell foundation models
Biological foundation models are increasingly judged by what their internal representations encode, not only by task accuracy.
5 preprints · 5 published · ↑ 1.9× vs 4-wk avg
Rank 3 Single-cell
Protocol and metric choices shape conclusions drawn from single-cell transcriptomics
Foundation and generative models now predict perturbation responses at scale, while protocol bias and metric choice still shape single-cell conclusions.
7 preprints · 3 published · ↑ 1.9× vs 4-wk avg
Rank 4 Tools & infrastructure
Reproducible tools for genotype prediction, PRS, and variant interpretation
Recent preprints and publications package prediction and variant-interpretation workflows as portable, self-contained software for genotype-based analysis.
5 preprints · 2 published
Rank 5 ML / AI methods
Protein language model embeddings drive ML prediction of protein function and interfaces
Proteome-scale ML predictors now offer linear-time interface prediction and screening across microbial and remote-homolog protein sets.
4 preprints · 2 published
Rank 6 Generative and optimization models design RNA, protein, peptide and drug candidatesStructural bio & protein designRank 7 Models and base editing link noncoding variants to regulatory element functionEpigenomicsRank 8 Graph models span drug repurposing, pocket representation, and toxicity predictionML / AI methodsRank 9 Comparative genomes link karyotype, gene-family and regulatory changes to adaptationGenomicsRank 10 Protein language models gain structural and chemical priors for prediction and designStructural bio & protein designRank 11 Foundation vs. task-specific deep models under limited biomedical dataML / AI methodsRank 12 Machine learning infers tumor features from histology, long reads, and cell-free DNAML / AI methodsRank 13 New algorithms for strain, eukaryotic, and long-read metagenomic analysisGenomicsRank 14 Benchmarking and statistical methods for detecting genomic errors and repeat variationGenomicsRank 15 Computational methods quantify spatial organization across tissues, neurons and speciesSingle-cellRank 16 Baseline-anchored validation exposes limits of learned protein structure predictorsStructural bio & protein designRank 17 Computational tools remove contaminant signal and resolve barcodes in sequencing dataTools & infrastructureRank 18 Learned models predict motifs and domains from protein sequence representationsML / AI methodsRank 19 Knowledge graphs structure biological associations for target ranking and annotationML / AI methodsRank 20 Refining population genomic inference of linkage, selection and Neanderthal introgressionGenomics
Show 20 more
Rank 21 Learned interface representations predict antibody, TCR-pMHC and virus-host bindingML / AI methodsRank 22 Perturbation models pursue generalization to unseen conditions with valid coverageSingle-cellRank 23 Generative models sample protein and genome conformations under learned physical priorsStructural bio & protein designRank 24 Learned latent models recover hidden cell-state structure from single-cell measurementsSingle-cellRank 25 Calibrated learned models refine protein-ligand affinity, screening, and dockingStructural bio & protein designRank 26 Simulations resolve conformational ensembles of flexible, aggregation-prone biomoleculesStructural bio & protein designRank 27 Scalable tools link sequencing data to variant calls, population structure and treesTools & infrastructureRank 28 Statistical tree estimation and tree-sampling methods advance phylogenomic inferenceGenomicsRank 29 Spectrum- and peptide-level scores drive microbial taxonomy and relatedness inferenceProteomicsRank 30 Evaluation-protocol sensitivity and attribution faithfulness in omics ML benchmarksSystems biologyRank 31 Annotation depth shapes transposable element inference across eukaryotic genomesGenomicsRank 32 Chromosome-scale assemblies for amoeba, hydrozoan, and annelid genomesGenomicsRank 33 Models of tumor-immune dynamics link macrophage reprogramming to treatment schedulingSystems biologyRank 34 Separating behavior-linked neural activity from internal dynamics across methodsSystems biologyRank 35 LLM agents constrained by fixed pipelines for single-cell metadata and model buildingSingle-cellRank 36 Interfaces and sequence rules govern condensate morphology and cross-β transitionsSystems biologyRank 37 Sequence composition and genome architecture as context-dependent, multi-scale variablesGenomicsRank 38 Assumption-light estimators for quantifying proteomic and fluorescence measurementsProteomicsRank 39 Benchmarking microbial genome references and pipelines for pseudogenes and phylogeniesTools & infrastructureRank 40 Quantum-accurate energies and charges via machine learning for biomolecular simulationStructural bio & protein design