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.
- Open access
- 1 cites
Moving in bioRxiv, Biotechnology Advances, Bioinformatics, Nature, Nature Methods, Molecular & Cellular Proteomics, Plant Physiology and Cell Systems.
Rebuilt
Nothing in this topic has moved yet this week.
Newest first · last 60 days
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
Golgi-endosomal CATCHR complexes assemble distinct trafficking modules, with COG, GARP, and EARP each preferentially associating with specific CCTs, Rabs, SNAREs, and SM proteins.
AI summary · may contain errors
BioLM-Score improves DEKOIS 2.0 screening endpoints over a matched geometry-only baseline, but affinity scoring and ranking do not improve uniformly under the same controlled comparisons.
AI summary · may contain errors
HieraPTM-HieraFuncPhos predicts functional phosphosites with biophysically grounded mechanical patterns at HSP90AA1 T36 and T115, matching published NMR observations.
AI summary · may contain errors
ARIES builds PLM-based multiple sequence alignments that achieve higher accuracy than state-of-the-art methods, especially in low-identity regimes, while scaling almost linearly with sequence count.
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
HandShake, a sequence-only partner-specific protein-protein interface predictor, achieves AUROC 0.811 and MCC 0.367 at 30% redundancy, exceeding the best prior sequence-only benchmark.
AI summary · may contain errors
A systematic review compares EC number prediction algorithms that infer enzyme reaction classes from reaction chemical structures, evaluating performance across databases and enzyme classes.
AI summary · may contain errors
SLAE-ERR predicts its own backbone 15N chemical-shift error with a mean absolute error of 0.76 ppm, using embeddings from SLAE trained on NMR-linked dynamic regions.
AI summary · may contain errors
DynamicsPLM, a 35-million-parameter dynamics-enriched protein language model, predicts zero-shot mutation effects from sequence by balancing evolutionary plausibility with structural dynamics disruption.
AI summary · may contain errors
Incorporating pretrained amino-acid sequence embeddings with mRNA-derived features improves translation efficiency prediction accuracy compared with using mRNA features alone.
AI summary · may contain errors
ModBERT, a DNA language model using a five-letter probabilistic genome representation including 5mC, predicts chromatin accessibility in held-out loci and individuals with AUROC 0.88.
AI summary · may contain errors
Region-aware pooling of mRNA embeddings (5′ UTR, coding sequence, 3′ UTR) improves linear-probe performance versus global mean pooling, boosting Pearson correlation by 11–25% on translation and stability tasks.
AI summary · may contain errors
ESM2 embeddings capture the sequence grammar of binary patterning in four-helix bundle designs, revealing a jellyfish-shaped manifold across 128 heptad patterns.
AI summary · may contain errors
SLiMMine, a deep learning method for the human proteome, identifies functional short linear motifs while eliminating ~80% of pattern-matching false-positive hits.
AI summary · may contain errors
DeePEn shows that protein variant-effect prediction performance steadily deteriorates as mutational distance from the wild type increases across four deep mutational scanning datasets.
AI summary · may contain errors
BIG-TB, built from over 17,000 Mycobacterium tuberculosis genomes with measured antibiotic resistance, shows DNA-based foundation models (mean test AUC 0.888) do not yet beat simple ML baselines (0.865).
AI summary · may contain errors
DELPHI, an open antibody developability prediction platform, achieves mean AUC 0.959 for polyreactivity and 0.933 for SEC using CDR H3-cluster cross-validation.
AI summary · may contain errors
mRNABench benchmarks mature mRNA property prediction on human transcripts using 13 datasets and 81 tasks, evaluating 29 nucleotide foundation model families across 750k experiments.
AI summary · may contain errors
A phylogenetic significance test for continuous traits finds amino-acid preferences linked to quantitative variation with higher recall at low false-positive rates, enabling whole-genome scans despite higher computation.
AI summary · may contain errors
SoluProtMut, a Siamese deep geometric neural network trained on deep mutational scanning in bacterial and yeast expression assays, achieves top informedness on 44 proteins, with a 26.5% ternary informedness for multipoint mutations.
AI summary · may contain errors
Stage 3 enables zero-shot residue-level protein-ligand binding-site prediction for previously unseen ligands, reaching pooled residue-level F1 0.3109 and ligand-macro F1 0.2338.
AI summary · may contain errors
In densely sampled combinatorial protein landscapes, protein language models show no consistent advantage over simple mutation-identity and evolutionary baselines for zero-shot scoring, supervised prediction, or library selection.
AI summary · may contain errors
structRFM, trained on 21 million RNA sequence–structure pairs, learns joint sequence and structure representations that rank highly in zero-shot homology classification and set new secondary-structure prediction benchmarks.
AI summary · may contain errors
Evo2 attains strong AUROC = 0.927 for mitochondrial variant-effect prediction, but shows systematic blind spots in short-range signals like codon usage bias and context sensitivity.
AI summary · may contain errors
DUET predicts RNA–small molecule binding sites with 0.240 +/- 0.013 MCC on the benchmark dataset, outperforming CoBRA-M3 at 0.192 +/- 0.031 under matched overlap controls.
AI summary · may contain errors
Computational prediction of protein-RNA interactions from 2010–2025 has evolved into five task categories, spanning from RBP classification to quantitative binding-affinity and mutation-effect estimation.
AI summary · may contain errors
AlphaFold’s series (AlphaFold2 to AlphaFold-Multimer to AlphaFold3) substantially advances peptide structure and multi-chain complex prediction using invariant point attention and ipTM confidence metrics.
AI summary · may contain errors
Integrated resources and multi-algorithm prediction tools for epigenetic regulation are expanding across DNA methylation, histone modification, RNA modification, ncRNAs, and disease-linked marks, with AI improving site prediction and functional mining.
Journals this month
Topics moving, week to 18 Nov 2026
By field