NACraft, a training-free and programmatic framework for all-atom nucleic-acid aptamer design based on backpropagation through structure-model feedback, is presented, demonstrating the effectiveness and versatility of NACraft and extending structure-model hallucination toward programmatic nucleic-acid aptamer design.
Abstract
Protein–nucleic-acid interactions underpin diverse biological processes and provide a basis for molecular sensing, regulation and therapeutic intervention. However, the coupled dependence of aptamer function on nucleotide sequence, three-dimensional folding and target binding makes rational RNA and DNA binder design challenging. Here we present NACraft, a training-free and programmatic framework for all-atom nucleic-acid aptamer design based on backpropagation through structure-model feedback. By composing binding, sequence-similarity and anti-binding constraints, NACraft supports de novo generation, similarity-guided sampling and target-selective design within a unified optimization framework, without task-specific training or fine-tuning. Computational experiments showed that NACraft generated high-confidence candidates de novo across diverse protein targets, with further improvements achieved through similarity-guided design for both RNA and DNA complexes. Its target-selective design capability was further validated in silico, with 69.44% of paired candidates generated to favour the positive target EGFR over the off-target HER2. Under matched independent AlphaFold3 evaluation, NACraft achieved better performance than ODesign in 10 of 11 NA-12 targets and 17 of 20 protein target–length settings. Together, these results demonstrate the effectiveness and versatility of NACraft and extend structure-model hallucination toward programmatic nucleic-acid aptamer design. Code https://github.com/OTEAM-AI4S/NACraft
HighPlay2 is presented as a feasible framework for the early-stage design and screening of cyclic peptide candidates containing ncAAs, while further affinity maturation and experimental structural validation remain necessary.
Huitian Lin, Wentong Wang, Ning Zhu et al.· European journal of medicina...· 0 citations
Cyclic peptides have emerged as a compelling class of bioactive scaffolds, but de novo design of target-binding cyclic peptides from protein structures remains challenging. Here, we present HighMorph, an interaction-guided framework that combines protein–protein interaction information with artificial intelligence for rational cyclic peptide design. HighMorph integrates Monte Carlo tree search with a Transformer-based policy-value network to efficiently explore cyclic peptide sequence space, while incorporating explicit atomic-level hydrogen bond constraints extracted from reference protein–protein complexes to guide sequence optimization. The framework is systematically validated on two clinically relevant targets, programmed death-ligand 1 (PD-L1) and kallikrein-related peptidase 4 (KLK4). Notably, 33.3% and 40% of the generated candidates are active against PD-L1 and KLK4, respectively, with active cyclic peptides exhibiting micromolar binding affinities (approximately 10–6 M). These results validate our approach for cyclic peptide design. Additionally, interaction analysis provides insights for developing therapeutics targeting challenging protein interfaces.
Minhui Lan, Chengyun Zhang, Wentong Wang et al.· Journal of Medicinal Chemist...· 0 citations
A machine-learning-assisted enzyme-engineering (MLEE) workflow that adds substrate-specific functional information to htFuncLib through an initial screening and sequencing round that may bypass the need for transition-state models and reduce the effort required for obtaining high-activity variants.
Li Wan, Mahdi Bagherpoor Helabad, Lena Fraedrich et al.· bioRxiv· 0 citations
Vilya-1 is introduced, a deep learning model that addresses two central challenges in macrocycle design: sampling biologically relevant conformations across arbitrary chemistries and predicting key developability properties such as membrane permeability.
Vilya Research Pascal Sturmfels, M. Salem, Naozumi Hiranuma et al.· 1 citation
DNA-targeting drugs exploit structural features of DNA, including base stacking, grooves, and noncanonical structures, to bind and modulate DNA function. Similarly, DNA aptamers leverage unique physicochemical environments created by diverse DNA conformations to achieve high-affinity recognition of small molecules, making them critical for biosensing applications. Despite their therapeutic and sensing potential, accurately predicting binding configurations within these dynamic structures remains a significant computational challenge requiring advanced molecular dynamics (MD) simulations powered by modern force fields. To evaluate modern AMBER-based parametrizations across diverse structural motifs, including aptamers, duplexes, and quadruplex-duplex hybrids, we performed dynamic docking simulations of five DNA-ligand pairs using multicanonical MD, a generalized-ensemble method, across five distinct force fields (OL15, OL21, OL24, bsc1, and tumuc1). Analysis of 750 μs of trajectory data revealed significant variations in conformational ensembles. OL24 exhibited the highest accuracy in reproducing experimental structures based on our R-value analysis, which quantifies the ligand-DNA native contacts, while OL15, OL21, and bsc1 also demonstrated robust performance across all systems. In contrast, tumuc1 displayed a persistent bias toward distorted or misoriented conformations with low native-state populations, compromising structural reliability regardless of the system type. These findings provide critical insights for developing next-generation DNA force fields capable of accurately modeling non-native structures and enabling balanced sampling essential for predicting ligand binding in diverse biological contexts.
G. Bekker, Y. Fukunishi, Junichi Higo et al.· Journal of Chemical Theory a...· 0 citations
RNA aptamers are often used as ligand-recognition modules in engineered RNA systems, but integration into larger RNA constructs can influence stability and ligand binding. As a result, aptamer sequences may need to be adapted to new environments while preserving essential properties. Here, we examine this sequence editability problem for the theophylline RNA aptamer. Starting from the experimentally determined structure, we introduced targeted mutations in peripheral structural elements while leaving the recognition site unchanged. The native aptamer, mutated variants, a Mg2+-depleted system, and a caffeine-bound control were analyzed using three independent 1 μs molecular dynamics simulations. Binding energetics were estimated with multiple end-point as well as alchemical free energy approaches. Results were interpreted together with base pair stability, the conformational landscape of the binding pocket, and per-nucleotide energy contributions. This allows us to predict whether an edit is tolerated or disruptive. Some mutations retained structural and energetic profiles comparable to the native aptamer, whereas others reduced ligand affinity by propagating structural distortions into the binding pocket. These results show that sequence changes outside the binding site can modulate ligand binding indirectly, and that the ligand interaction network is useful for evaluating edited aptamers. The introduced workflow provides a novel combination of established computational strategies for efficient in silico screening of aptamer variants before experimental testing and can be integrated into the design of larger RNA structures. This works particularly well when an experimental structure is available and the tested mutations are small enough not to disrupt the folding pathway.
Aamir Aman, Leonhard Sidl, Nitchakan Darai et al.· International Journal of Mol...· 0 citations