DeCAF Framework Enables Fast All-Atom Protein Structure Prediction with Few Inference Steps
Researchers introduced DeCAF, a framework that distills state-of-the-art all-atom cofolding models into flow maps capable of generating high-quality protein structures in only a few inference steps rather than expensive iterative processes. The method builds on denoiser-based flow maps with endpoint losses supporting SE(3) rigid alignment and operates in the noise schedule of EDM-style architectures. This advancement matters because it significantly reduces computational costs for protein structure prediction while maintaining or improving accuracy, making deployment and inference-time search more practical.
The DeCAF (Denoiser Cofolding All-Atom Flowmap) framework addresses a key computational bottleneck in all-atom generative modeling of biomolecular complexes by distilling expensive diffusion-based cofolding models into efficient flow maps. The approach uses a denoiser-based formulation with endpoint losses that naturally support SE(3) rigid alignment, critical for training accurate models, and employs a change of variables enabling direct distillation from pretrained diffusion models. Empirical results show DeCAF-Boltz statistically improves over Boltz-1x in both RMSD accuracy and physical validity scores on protein-ligand pose prediction, while DeCAF-Pearl matches its teacher model's success rate while using 5 times fewer network function evaluations. The framework also introduces a reward-guided search mechanism for inference-time sampling improvement, demonstrating a more optimal Pareto frontier across different computational budgets on benchmark datasets.
What's missing
The study does not discuss potential limitations of the flow map distillation approach, such as whether performance gains hold across diverse protein families or under different experimental conditions, nor does it address computational requirements for the distillation process itself or generalization to other biomolecular systems beyond protein-ligand complexes.
What different sources said
- arXiv cs.LGCenter
Few-step Cofolding with All-Atom Flow Maps
Related
Gut Bacteria Enzyme Found to Break Down Heat-Processed Food Compounds, Producing Novel Biogenic Amines
Researchers have discovered that an enzyme in common gut bacteria can degrade N-epsilon-carboxymethyllysine (CML), a compound formed during thermal food processing, producing previously unknown biogenic amines. The enzyme, ornithine decarboxylase SpeC from enterobacteria, acts on CML and related modified lysine derivatives through a low-level 'underground' catalytic activity. This finding suggests a previously unrecognized communication axis between thermally processed dietary compounds and gut microbial physiology, with potential implications for host health.
Full-Length Gene Sequencing Reveals Two Distinct Bacterial Communities in Black-Legged Ticks Expanding Into Canada
Researchers used Oxford Nanopore full-length 16S rRNA gene sequencing to characterize the microbiome of Ixodes scapularis black-legged ticks collected in Nova Scotia, Canada, distinguishing between tick-adapted bacteria and environmentally acquired bacteria. The study comes as I. scapularis — the primary vector of Lyme disease — is rapidly expanding northward into Canada due to climate change. The findings suggest that environmentally derived bacteria in tick microbiomes are not mere contamination, which has implications for how tick microbiome data is collected and interpreted across surveillance studies.
Study Identifies Metabolic Link Between Cell Envelope Stress and Biofilm Formation in Bacteria
Researchers have discovered that the metabolite acetyl-CoA directly inhibits enzymes that degrade the bacterial signaling molecule c-di-GMP, connecting cell envelope biosynthesis stress to biofilm formation in Pseudomonas aeruginosa. The study found that sub-inhibitory concentrations of antibiotics targeting early peptidoglycan biosynthesis — but not other antibiotic classes — elevate c-di-GMP levels by reducing phosphodiesterase activity, with acetyl-CoA competing for the enzyme active site. Because the relevant enzyme domain is broadly conserved across bacterial species, this checkpoint mechanism may be widespread and could have implications for understanding antibiotic-induced biofilm responses.