TrendBio

Top 40

View all ›

BACKFILLED Code and industry signals are partial for backfilled days.

Rank 1

ML / AI methods

Graph and transformer models integrate spatial and single-cell multi-omics data

Population-scale spatial and single-cell datasets across modalities are driving learned models that add uncertainty estimates and interpretability.

8 preprints · 5 published · ↑ 4.3× vs 4-wk avg
Rank 2

Structural bio & protein design

Generative nucleotide design couples learned models to structural and regulatory…

Recent preprints and papers pair learned generators with 3D structure, folding context, host codon usage, or cell-type specificity for nucleic acid design.

7 preprints · 2 published · ↑ 3.2× vs 4-wk avg
Rank 3

Single-cell

Isoform-resolved single-cell and single-molecule methods quantify transcript variants

Long-read and 5' single-cell profiling now resolve isoform and start-site usage per cell, with degradation-corrected quantification.

4 preprints · 4 published · ↑ 2.9× vs 4-wk avg
Rank 4

Genomics

Pangenome and graph references are improving structural variant calling accuracy

Pangenome graphs, long-read assemblies, and haplotype-resolved genomes are now used to improve structural variant detection accuracy.

3 preprints · 3 published · ↑ 3.0× vs 4-wk avg
Rank 5

Structural bio & protein design

Physics-based sampling and free energy methods for protein-ligand and protein-protein…

Direct side-by-side benchmarks of binding free energy methods and coarse-grained sampling in docking are now emerging across preprints and publications.

5 preprints · 3 published · ↑ 2.9× vs 4-wk avg
Rank 6 Graph and language-model encoders advance drug-target and synergy prediction↑ 2.2×ML / AI methodsRank 7 Fitted mechanistic models link single-cell and patient omics to regulatory dynamics↑ 2.3×Single-cellRank 8 Deep learning predicts RNA-binding sites, lncRNAs, and cancer epigenetic regulators↑ 2.3×ML / AI methodsRank 9 Open computational tools for bacterial recombination, community design, and taxonomic…↑ 2.1×Tools & infrastructureRank 10 LLM agents and multimodal models convert literature and spectra into molecular knowledge↑ 2.1×ML / AI methodsRank 11 Comparative genomes link repeat activity and duplication to lineage-specific gene loss↑ 2.0×GenomicsRank 12 Spatial models couple influenza dispersal, defective particles and interferon signalling↑ 1.9×ML / AI methodsRank 13 Reduced recurrent models reproduce complex dynamics across neural and molecular systems↑ 2.0×Systems biologyRank 14 Learned models predict one genomic readout from another across modalities↑ 1.8×Systems biologyRank 15 Computational methods refine single-cell and spatial transcriptomic measurements↑ 1.8×Single-cellRank 16 Ancient-DNA and structural-variation methods sharpen tests of selection in human genomes↑ 1.9×GenomicsRank 17 Integrated variant and annotation models improve polygenic and genomic prediction↑ 1.6×GenomicsRank 18 Benchmarking genomic and protein prediction models with reusable longitudinal evaluation↑ 1.6×GenomicsRank 19 Multi-omic and structural data are converted into drug and target prioritization outputs↑ 1.6×Systems biologyRank 20 Benchmarks and packaged pipelines standardize omics analysis tools↑ 1.5×Tools & infrastructure