The growing support for noncanonical amino acids in structure prediction tools such as AlphaFold3 has been largely facilitated by the Chemical Component Dictionary (CCD). However, the limited coverage of modified residues in CCD continues to restrict the application of these models to many biologically and therapeutically relevant peptides. To address this gap, we present HighRes_Builder, a computational method for efficient residue search and automated construction of noncanonical amino acids not currently archived in CCD. We demonstrate the utility of our approach by predicting structures for 3179 noncanonical residues beyond the CCD using AlphaFold3. AlphaFold3 achieved 100% acceptance for both the noncanonical residue monomers and their corresponding 'GGXGG' motifs (where X denotes the noncanonical residue). Of these, 72.44% of the predicted residue monomer structures concurrently satisfy all five geometric criteria (d_N_C1, d_Ck_Ccarb, d_Ccarb_O_mean, ang_Ck_Ccarb_O_mean, and ang_O1_Ccarb_O2). Furthermore, among the generated motif structures, 85.78% exhibited favorable ω values for the embedded noncanonical residues. Furthermore, by integrating HighRes_Builder with structure prediction systems, AlphaFold3 for linear peptides and HighFold3 for cyclic peptides, we successfully model the conformation of the linear peptide drug Relamorelin and the cyclic therapeutic peptides LUNA18 and JNJ-77242113 in complex with their target proteins, elucidating structural determinants of their mechanism of action. This work establishes a scalable and accurate framework for structure prediction of diverse nonstandard peptides, highlighting its potential to accelerate rational design of peptide-based therapeutics.

Drug resistance is a significant challenge in cancer treatment, greatly impacting treatment efficacy. Non-coding RNAs (ncRNAs) play crucial roles in mediating drug resistance, yet few computational models effectively predict drug resistance-associated ncRNAs. Existing methods often overlook the complex sequence patterns of ncRNA and their intricate interrelationships, resulting in suboptimal performance. To address these challenges, we propose MIFNDRA, a multimodal integrative framework that jointly models ncRNA and drug features to identify drug resistance-related ncRNAs. MIFNDRA employs a pre-trained Graph Isomorphism Network to extract drug structural features and a pre-trained SpliceBERT model to encode ncRNA sequences. It also incorporates various similarity features for both drugs and ncRNAs, while improving representation through a novel ncRNA interaction network that includes interactions between different ncRNA types as a strategy for knowledge enhancement. By leveraging advanced graph learning techniques, including residual GraphSAGE and contrastive learning, the model improves the identification of drug resistance-associated ncRNAs. Additionally, we curated a new benchmark dataset pairing ncRNA sequences with drug SMILES and resistance annotations. Comprehensive experiments demonstrate that MIFNDRA achieved state-of-the-art performance. Case studies on cisplatin and gemcitabine further validate the model's robustness and potential in advancing drug resistance research and drug development. The data and code required for this work are available at https://github.com/SJNNNN/MIFNDRA.

Cyclic peptides become attractive therapeutic candidates due to their diverse biological activities. However, existing deep learning-based sequence design models, such as ProteinMPNN, are primarily intended for linear peptides or proteins and do not explicitly account for the unique topological constraints of cyclic peptides. In this study, we introduce HighMPNN, a graph neural network model specifically developed for cyclic peptide sequence design. Through the integration of explicit structural constraints into the GNN-based framework, HighMPNN captures the geometric features of cyclic backbones while learning sequence patterns. The combination of cross-entropy loss with Frame Aligned Point Error (FAPE) loss allows the model to simultaneously optimize sequence generation and enhance structural accuracy. HighMPNN demonstrates superior performance in both sequence recovery rate and structural consistency compared to baseline models, achieving an average sequence recovery rate of 63.95% and an average Cα root-mean-square deviation (RMSD_Cα) of 1.413 Å. These results highlight the model's ability to generate sequences that closely resemble native backbones. At present, HighMPNN is limited to natural amino acids. Future work will focus on extending the framework to support non-canonical residues and structurally diverse cyclic peptide scaffolds, thereby accelerating cyclic peptide discovery and advancing peptide-based drug development.

Cyclic peptides have emerged as a research hotspot in drug development in recent years due to their excellent stability, specificity, and cell penetration. However, existing computational models face challenges in accurately predicting the three-dimensional structures of cyclic peptides containing unnatural amino acids (unAAs), thereby limiting their drug design. The release of AlphaFold 3 has significantly enhanced the modeling capability of biomolecular complexes and enabled the inclusion of unAAs through definitions provided by the Chemical Component Dictionary (CCD). Nevertheless, its training data reliance limits its ability to accurately predict cyclic peptide structures, failing to meet the demand for precise cyclic peptide structure prediction. Based on the AlphaFold 3 framework, we developed HighFold3 by introducing the Cyclic Position Offset Encoding Matrix (CycPOEM). HighFold3 comprises two submodels: HighFold3-Linear and HighFold3-Cyclic, designed for predicting the structures of linear and cyclic peptides, respectively. Our results demonstrate that HighFold3 outperforms existing models (HighFold, HighFold2, CyclicBoltz1, NCPepFold, CABS-flex, ESMFold, and HelixFold) in cyclic peptide structure prediction. It achieves atomic-level precision in predicting cyclic peptide monomers while demonstrating enhanced accuracy and generalization capability for cyclic peptide complexes containing unAAs. This offers unprecedented technical support for the structural design and optimization of cyclic peptide-based therapeutics.

Understanding the relationship between protein sequences and structures is essential for accurate protein property prediction. We propose BridgeNet, a pre-trained deep learning framework that integrates sequence and structural information through a novel latent environment matrix, enabling seamless alignment of these two modalities. The model's modular architecture-comprising sequence encoding, structural encoding, and a bridge module-effectively captures complementary features without requiring explicit structural inputs during inference. Extensive evaluations on tasks such as enzyme classification, Gene Ontology annotation, coenzyme specificity prediction, and peptide toxicity prediction demonstrate its superior performance over state-of-the-art models. BridgeNet provides a scalable and robust solution, advancing protein representation learning and enabling applications in computational biology and structural bioinformatics.

Cyclic peptides containing unnatural amino acids possess many excellent properties and have become promising candidates in drug discovery. Therefore, accurately predicting the 3D structures of cyclic peptides containing unnatural residues will significantly advance the development of cyclic peptide-based therapeutics. Although deep learning-based structural prediction models have made tremendous progress, these models still cannot predict the structures of cyclic peptides containing unnatural amino acids. To address this gap, we introduce a novel model, HighFold2, built upon the AlphaFold-Multimer framework. HighFold2 first extends the pre-defined rigid groups and their initial atomic coordinates from natural amino acids to unnatural amino acids, thus enabling structural prediction for these residues. Then, it incorporates an additional neural network to characterize the atom-level features of peptides, allowing for multi-scale modeling of peptide molecules while enabling the distinction between various unnatural amino acids. Besides, HighFold2 constructs a relative position encoding matrix for cyclic peptides based on different cyclization constraints. Except for training using spatial structures with unnatural amino acids, HighFold2 also parameterizes the unnatural amino acids to relax the predicted structure by energy minimization for clash elimination. Extensive empirical experiments demonstrate that HighFold2 can accurately predict the 3D structures of cyclic peptide monomers containing unnatural amino acids and their complexes with proteins, with the median RMSD for Cα reaching 1.891 Å. All these results indicate the effectiveness of HighFold2, representing a significant advancement in cyclic peptide-based drug discovery.

In recent years, cyclic peptides have emerged as a promising therapeutic modality due to their diverse biological activities. Understanding the structures of these cyclic peptides and their complexes is crucial for unlocking invaluable insights about protein target-cyclic peptide interaction, which can facilitate the development of novel-related drugs. However, conducting experimental observations is time-consuming and expensive. Computer-aided drug design methods are not practical enough in real-world applications. To tackles this challenge, we introduce HighFold, an AlphaFold-derived model in this study. By integrating specific details about the head-to-tail circle and disulfide bridge structures, the HighFold model can accurately predict the structures of cyclic peptides and their complexes. Our model demonstrates superior predictive performance compared to other existing approaches, representing a significant advancement in structure-activity research. The HighFold model is openly accessible at https://github.com/hongliangduan/HighFold.

Protein-peptide interactions (PPepIs) are vital to understanding cellular functions, which can facilitate the design of novel drugs. As an essential component in forming a PPepI, protein-peptide binding sites are the basis for understanding the mechanisms involved in PPepIs. Therefore, accurately identifying protein-peptide binding sites becomes a critical task. The traditional experimental methods for researching these binding sites are labor-intensive and time-consuming, and some computational tools have been invented to supplement it. However, these computational tools have limitations in generality or accuracy due to the need for ligand information, complex feature construction, or their reliance on modeling based on amino acid residues. To deal with the drawbacks of these computational algorithms, we describe a geometric attention-based network for peptide binding site identification (GAPS) in this work. The proposed model utilizes geometric feature engineering to construct atom representations and incorporates multiple attention mechanisms to update relevant biological features. In addition, the transfer learning strategy is implemented for leveraging the protein-protein binding sites information to enhance the protein-peptide binding sites recognition capability, taking into account the common structure and biological bias between proteins and peptides. Consequently, GAPS demonstrates the state-of-the-art performance and excellent robustness in this task. Moreover, our model exhibits exceptional performance across several extended experiments including predicting the apo protein-peptide, protein-cyclic peptide and the AlphaFold-predicted protein-peptide binding sites. These results confirm that the GAPS model is a powerful, versatile, stable method suitable for diverse binding site predictions.

MOTIVATION: 5-Methylcytosine (5mC), a fundamental element of DNA methylation in eukaryotes, plays a vital role in gene expression regulation, embryonic development, and other biological processes. Although several computational methods have been proposed for detecting the base modifications in DNA like 5mC sites from Nanopore sequencing data, they face challenges including sensitivity to noise, and ignoring the imbalanced distribution of methylation sites in real-world scenarios. RESULTS: Here, we develop NanoCon, a deep hybrid network coupled with contrastive learning strategy to detect 5mC methylation sites from Nanopore reads. In particular, we adopted a contrastive learning module to alleviate the issues caused by imbalanced data distribution in nanopore sequencing, offering a more accurate and robust detection of 5mC sites. Evaluation results demonstrate that NanoCon outperforms existing methods, highlighting its potential as a valuable tool in genomic sequencing and methylation prediction. In addition, we also verified the effectiveness of our representation learning ability on two datasets by visualizing the dimension reduction of the features of methylation and nonmethylation sites from our NanoCon. Furthermore, cross-species and cross-5mC methylation motifs experiments indicated the robustness and the ability to perform transfer learning of our model. We hope this work can contribute to the community by providing a powerful and reliable solution for 5mC site detection in genomic studies. AVAILABILITY AND IMPLEMENTATION: The project code is available at https://github.com/Challis-yin/NanoCon.