AI summary · may contain errors
PlantCAD2, a plant-specific DNA language model with 676 million parameters, captures evolutionary conservation and surpasses Evo2 in 10 of 12 zero-shot tasks across 65 angiosperm genomes.
- 28 cites
Moving in Cell Genomics.
Rebuilt
AI summary · may contain errors
PlantCAD2, a plant-specific DNA language model with 676 million parameters, captures evolutionary conservation and surpasses Evo2 in 10 of 12 zero-shot tasks across 65 angiosperm genomes.
Newest first · last 60 days
AI summary · may contain errors
CADD-SV v2.0 delivers a unified Random Forest framework that scores deleteriousness of deletions, insertions, duplications, and inversions using both coordinates and flanking sequence context.
AI summary · may contain errors
On coding sequences of 3,244 human genes with a homology-aware split, the frozen encoder NT-v2 recovers coarse protein-family signal only at about nucleotide-composition level (macro-F1 0.693 vs 0.639).
AI summary · may contain errors
A graph-based pangenome for Cucurbita pepo from nine chromosome-level assemblies identified 229,431 high-confidence structural variants, genotyped across 206 wild and cultivated accessions.
AI summary · may contain errors
Codon usage bias across 262 plant genomes is strongly non-random, phylogenetically structured, and shows progressive relaxation from algae to angiosperms while conserving a ~30 ATP per-codon economy cost.
AI summary · may contain errors
An extended clade framework extending the conditional clade distribution (CCD) model improves summary of annotated Bayesian phylogenetic trees by explicitly using node annotations.
AI summary · may contain errors
TRIDENT, an automated eDNA metabarcoding assignment workflow for marine vertebrates, integrates molecular similarity, curated taxonomy, and ecological plausibility to improve taxonomic assignments.
AI summary · may contain errors
reconcILS reconciles gene trees to a species tree while accounting for incomplete lineage sorting by modeling ILS as nearest neighbor interchange events, improving accuracy.
AI summary · may contain errors
In a DIA-NN proteomics dataset, run-wise q-value filtering at nominal 1% FDR yields a recomputed decoy ratio of 2.2% with 0.93 Jaccard overlap, and added identifications show a twelvefold higher entrapment-match rate.
AI summary · may contain errors
Seven structural classes (A–G) of prokaryotic DNA methyltransferases were identified by catalytic-domain 3D topology, using 38 experimentally determined structures plus AlphaFold models.
AI summary · may contain errors
In 639 isolate pairs of Enterococcus faecium, Klebsiella pneumoniae, Listeria monocytogenes, and Staphylococcus aureus sequenced on one nanopore platform, 40.37% of read placements between strains above 99% ANI show no sequence difference.
AI summary · may contain errors
GradeBins accurately measures quality of prokaryotic metagenome bins, matching independent calculations for 6,374 simulated bins and achieving the highest observed F1 for bin selection with 20 kb fragments.
AI summary · may contain errors
Clair3-RNA showed the broadest, most robust small-variant calling applicability across ONT and PacBio lrRNA-seq chemistries, outperforming no single caller overall but spanning contexts.
AI summary · may contain errors
A BERT-tokenized CNN-BiLSTM model predicts PCR amplification outcomes with 85.9% accuracy by encoding primer-template match and mismatch interactions at Gibbs free energy–defined binding sites.
AI summary · may contain errors
ISdetector maps insertion sequences and linked structural variants from bacterial short reads, achieving a high-confidence F1 score of 0.86 on 19 Mycobacterium tuberculosis genomes.
AI summary · may contain errors
Lineage-specific substitution rate variation severely biases phylogenetic network inference, inflating find_graphs worst f-statistic residuals beyond the 3 threshold and raising type I error for PhyNEST.
AI summary · may contain errors
Toxin genes in four venomous bilaterian species are tandem-arrayed beyond gene-family-size expectations, with clustering enrichment dropping from 14–31-fold to about 2–6-fold versus size-matched non-toxin families.
AI summary · may contain errors
Qitan Technology’s QNome-3841 with QCell-384 recovered valid arthropod DNA barcodes from 98.5% of specimens, nearly matching Oxford Nanopore, but with lower per-read identity (98.04% vs 99.39%).
AI summary · may contain errors
Homo sapiens Papuans show Denisovan DNA related to immune function, recovered using the D* statistic with allelic rarefaction.
AI summary · may contain errors
A chromosome-scale genome assembly and multi-omic atlas for the common bent-wing bat Miniopterus schreibersii integrates lncRNA, EVE mapping, and same-individual RNA editomics to enable immune-regulation studies.
AI summary · may contain errors
hilldiv3 unifies neutral, phylogenetic, and functional Hill-number diversity analysis in R, supporting all twelve benchmarked operation–type combinations and building pairwise (dis)similarity matrices 15–820× faster than hilldiv2 and hillR.
AI summary · may contain errors
MitoScope, an end-to-end long-read mitochondrial genome workflow, achieves high sensitivity and precision on SMaHT Network benchmarking while reducing artifacts from repeats and NUMTs.
AI summary · may contain errors
MOTET jointly infers multimodal mutational processes in 2,583 PCAWG whole genomes, recovering known MPs more accurately than standard post hoc correlation methods.
AI summary · may contain errors
Across 60 eukaryote groups, published species-level phylogenies include 19.0% of described species, with only 7.5% placed by molecular data and 5.4% in dated trees.
AI summary · may contain errors
Trio genome sequencing in 108 children from 100 families with complex neuromotor phenotypes identified genetic diseases explaining the phenotype in 53 families.
AI summary · may contain errors
MicroDUET, a dual-headed ProkBERT-based genomic foundation model, identifies pathogen species, gene families, and resistance-linked function with up to 97.3% accuracy and AUROC 0.985.
AI summary · may contain errors
Chimera aligns single-cell atlases across species and finds that ortholog expression deployment evidence ({Delta}[a]) can diverge from protein-sequence conservation, including in held-out species.
AI summary · may contain errors
In 141 Escherichia coli adaptive laboratory evolution experiments, Fisher exact testing at 5% FDR found 74 significant gene pairs (36 co-occurrences, 38 mutual exclusivities).
AI summary · may contain errors
A chromosome-level genome assembly for Dracocephalum ruyschiana resolves two pseudo-haplotypes (661 Mb and 681 Mb) scaffolded into seven chromosomes each.
AI summary · may contain errors
The last bacterial common ancestor encoded a functionally complete flagellum, with 50 ancestral protein groups reconstructed from 80,789 bacterial proteomes.
Journals this month
Topics moving, week to 18 Nov 2026
By field