Biology subjects

Roy, R. S.

Publications and source records attributed to Roy, R. S..

13 recordsLinked to original sources

bioRxiv2026

CARTnet: Template-guided B-cell Epitope Prediction with Contact-aware Attention

B-cell epitope prediction remains challenging due to the limited availability of experimentally characterized data and the incomplete understanding of antibody-antigen recognition. Here, we present CARTnet, a novel contact-aware deep learning framework that integrates multimodal features and template-derived epitope information. CARTnet demonstrated superior performance when compared to the existing state-of-the-art methods across most evaluation metrics on both the internal and external test datasets. In particular, CARTnet achieved an AUPRC of 0.243 on the internal test set, compared with 0.220 for DiscoTope-3, 0.140 for BepiPred-2, and 0.198 for BepiPred-3. This improvement was also observed on the external test set, where CARTnet achieved an AUPRC of 0.267, compared with 0.192, 0.111, and 0.215 for DiscoTope-3, BepiPred-2, and BepiPred-3, respectively, supporting its generalizability. CARTnet also supports template-guided prediction, in which limited prior epitope information can improve the identification of additional epitope residues; a single template residue increased the AUPRC of the template-based variant from 0.225 to 0.403 on the internal test set. Furthermore, we demonstrated that epitope information from closely related viruses can be transferred across the genus Alphavirus to improve epitope prediction accuracy. These results establish CARTnet as an accurate framework for both de novo and template-guided B-cell epitope prediction.

bioinformatics↗

bioRxiv2026

Agent K: An Agent-Orchestrated Pipeline for Mapping Genes to Biosynthetic Pathway towards Metabolite Discovery

The identification of biosynthetic gene clusters (BGCs) and their associated metabolic path- ways is central to genome-guided natural product discovery. However, existing computational tools depend on fragmented workflows for data processing, annotation, and analysis, which re- quire substantial manual intervention and lack seamless integration across heterogeneous plat- forms and tools. Here, we present, Agent K, an agent-orchestrated framework that automates the end-to-end process from bacterial whole genome sequence analysis to biosynthetic pathway mapping and metabolite discovery as candidate gene products for functional insights. The sys- tem employs a large-language-model (LLM) agent to coordinate multiple bioinformatics tools, enabling dynamic execution, data integration, result interpretation, and flexible and intuitive visualization. By integrating genomic sequences to functional annotations and pathway-level reconstructions, the framework facilitates the identification of candidate gene products, their biosynthetic pathways associated with individual BGCs, and the putative metabolites produced by those pathways. The developed novel agentic framework provides a user-friendly scalable and extensible solution for automated high-throughput analysis and mapping of bacterial secondary metabolism pathways for interactive exploration of candidate potential.

bioinformatics↗

bioRxiv2026

An Integrated Computational Antigen Discovery Pipeline with Hierarchical Filtering for Emerging Viral Variants

Emerging and evolving viral diseases, such as SARS-CoV-2, continue to pose significant global health challenges, underscoring the urgent need for rapid and scalable antigen discovery pipelines. This work presents a computational pipeline that integrates diverse computational tools and machine learning models to accelerate the identification and optimization of antigen candidates. The pipeline employs efficient filtering and consensus-based strategies to highlight epitopes with high therapeutic potential. We demonstrate its utility by significantly narrowing the antigen search space for Rift Valley fever virus (RVFV) and Mayaro virus (MAYV), and by effectively identifying conserved neutralizing epitopes in SARS-CoV-2. Our proposed computational antigen pipeline offers a powerful framework for expediting the development of future vaccines and therapeutics in response to emerging pathogens.

bioinformatics↗

bioRxiv2026

Structure-Based TCR-pMHC Binding Prediction and Generalization to Unseen Peptides

The interaction between T-cell receptors (TCRs) with the peptide-bound major histocompatibility complex (MHC) intricately impacts the functional specificity of T-cell-mediated adaptive immune response. Consequently, implication in immunotherapy has contributed to the ever-growing computational methods for TCR recognition, which have recently attracted structure-based approaches due to advancements in protein structure modeling. Despite access to structural information of the predicted binding interface, graph neural network (GNN)-based TCR-pMHC binding specificity classifiers tend to show poor accuracy for samples with unseen peptides. In this work, we comprehensively assess the potential factors that critically impact the generalization performance of classifiers trained with computationally predicted structures. Specifically, our experiments focus on analyzing the sensitivity of such predictors to the interaction features in the TCR-pMHC interface and the structural uncertainty. Building on the analysis, we demonstrate how the design of classifier architecture with auxiliary training objectives can improve the generalization performance to novel peptides not yet seen during model training. Overall, our work highlights the challenges of unseen peptide generalization from different perspectives of the GNN-based classifier paradigm, showcasing the strengths and weaknesses of the current state-of-the-art approaches in the generalization landscape.

bioinformatics↗

bioRxiv2025

Pathogenic DVL frameshifting variants in Robinow syndrome disrupt WNT signaling and cellular dynamics

Robinow syndrome (RS) is a genetically heterogeneous rare disorder involving six genes in the WNT/planar cell polarity (PCP) signaling pathway. Frameshifting variants in DVL genes that introduce a novel basic C-terminus are a common cause of autosomal dominant RS, accounting for [~]33% of individuals without ROR2 variants. Here, we review ClinVar and literature variants affecting DVL paralogs resulting in RS and investigate the cellular effects of pathogenic DVL frameshift variants with mutant tails replacing at least 82 amino acids. In silico analysis of the new C-termini revealed altered intrinsically disordered regions (IDRs), charge distribution, and predicted protein structures. To explore potential altered biological mechanisms caused by novel C-termini, we generated wild-type (WT), frameshift, and truncated constructs of DVL1-3, and analyzed their behavior in a transfection-based in vitro systems. DVL proteins normally polymerize into cytoplasmic puncta that redistribute upon WNT stimulation. Immunocytochemistry showed that mutant DVL proteins failed to change their localization in response to WNT ligands, in contrast to WT alleles--a consistent observation across all three DVLs. In line with this, TOPFlash reporter assays demonstrated that mutant DVL1 and DVL3 failed to activate canonical WNT signaling, while WT proteins induced strong activation. Additionally, the mutant C-terminal tail interfered with CSNK1E-induced phosphorylation, offering a potential mechanism underlying the impaired WNT response. Our results provide further understanding of the cellular consequences of pathogenic DVL frameshifting variants and offers insights into the effects of such alleles on WNT signaling, and the perturbations thereof, that may lead to developmental phenotypes observed in RS.

genetics↗

bioRxiv2025

Axonal defasciculation is restricted to specific branching points during regeneration of the lateral line nerve in zebrafish

Peripheral nerve regeneration requires precise selection of the appropriate targets of innervation, often in an environment that differs from that during the developmental wiring of the neural circuit. Severed axons of the zebrafish posterior lateral line nerve have the capacity to reinnervate mechanosensory hair cells clustered in neuromast organs. Regeneration represents a balance between fasciculated regrowth of the axonal bundle and defasciculation of individual axons into the epidermis where neuromasts reside. The cues that guide pathfinding during regeneration of the posterior lateral line nerve are unknown. Here we show that expression of col18a1a, which codes for the secreted heparan sulfate proteoglycan collagen XVIII, biases axonal defasciculation to specific branching points that coincide with circumscribed gaps in the epidermal boundary. We found that col18a1a is expressed by the neuromast and by a subset of Schwann cells that are located at the points of axonal defasciculation. Furthermore, we observed axon branching at inappropriate locations during nerve regeneration in col18a1a mutants. We propose a model in which a collagen XVIII-based axon-guidance cue complex attracts defasciculated axons across the epidermal basement membrane. Summary StatementThe success of nerve regeneration depends on precise axon pathfinding and accurate target selection. We identify neuron-extrinsic factors that guide regeneration in a zebrafish model.

neuroscience↗

bioRxiv2023

Enhancing AlphaFold-Multimer-based Protein Complex Structure Prediction with MULTICOM in CASP15

AlphaFold-Multimer has emerged as the state-of-the-art tool for predicting the quaternary structure of protein complexes (assemblies or multimers) since its release in 2021. To further enhance the AlphaFold-Multimer-based complex structure prediction, we developed a new quaternary structure prediction system (MULTICOM) to improve the input fed to AlphaFold-Multimer and evaluate and refine the outputs generated by AlphaFold2-Multimer. Specifically, MULTICOM samples diverse multiple sequence alignments (MSAs) and templates for AlphaFold-Multimer to generate structural models by using both traditional sequence alignments and new Foldseek-based structure alignments, ranks structural models through multiple complementary metrics, and refines the structural models via a Foldseek structure alignment-based refinement method. The MULTICOM system with different implementations was blindly tested in the assembly structure prediction in the 15th Critical Assessment of Techniques for Protein Structure Prediction (CASP15) in 2022 as both server and human predictors. Our server (MULTICOM_qa) ranked 3rd among 26 CASP15 server predictors and our human predictor (MULTICOM_human) ranked 7th among 87 CASP15 server and human predictors. The average TM-score of the first models predicted by MULTICOM_qa for CASP15 assembly targets is [~]0.76, 5.3% higher than [~]0.72 of the standard AlphaFold-Multimer. The average TM-score of the best of top 5 models predicted by MULTICOM_qa is [~]0.80, about 8% higher than [~]0.74 of the standard AlphaFold-Multimer. Moreover, the novel Foldseek Structure Alignment-based Model Generation (FSAMG) method based on AlphaFold-Multimer outperforms the widely used sequence alignment-based model generation. The source code of MULTICOM is available at: https://github.com/BioinfoMachineLearning/MULTICOM3.

bioinformatics↗

bioRxiv2023

Improving AlphaFold2-based Protein Tertiary Structure Prediction with MULTICOM in CASP15

Since CASP14, AlphaFold2 has become the standard method for protein tertiary structure prediction. One remaining challenge in the field is to further improve the accuracy of AlphaFold2-based protein structure prediction. To address this challenge, we developed a new version of the MULTICOM system to sample diverse multiple sequence alignments (MSAs) and structural templates to improve the input for AlphaFold2 to generate structural models. The models are then ranked by both the pairwise model similarity and AlphaFold2 self-reported model quality score. The top ranked models are further refined by a novel structure alignment-based refinement method powered by Foldseek. Moreover, for a monomer target that is a subunit of a protein assembly (complex), MULTICOM integrates tertiary and quaternary structure prediction together to account for tertiary structural changes induced by protein-protein interaction in the assembly. The MULTICOM system participated in the tertiary structure prediction in the 15th Critical Assessment of Techniques for Protein Structure Prediction (CASP15) in 2022 as server and human predictors. Our best server predictor (MULTICOM_refine) ranked 3rd among 47 CASP15 server predictors and our best human predictor (MULTICOM) ranked 7th among all 132 human and server predictors. The average GDT-TS score and TM-score of the first structural models that MULTICOM_refine predicted for 94 CASP15 domains are [~]0.80 and [~]0.92, 9.6% and 8.2% and higher than [~]0.73 and 0.85 of the standard AlphaFold2 predictor respectively. The results demonstrate that our approach can significantly improve the accuracy of the AlphaFold2-based protein tertiary structure prediction. The source code of MULTICOM is available at: https://github.com/BioinfoMachineLearning/MULTICOM3.

bioinformatics↗

bioRxiv2023

Combining pairwise structural similarity and deep learning interface contact prediction to estimate protein complex model accuracy in CASP15

Estimating the accuracy of quaternary structural models of protein complexes and assemblies (EMA) is important for predicting quaternary structures and applying them to studying protein function and interaction. The pairwise similarity between structural models is proven useful for estimating the quality of protein tertiary structural models, but it has been rarely applied to predicting the quality of quaternary structural models. Moreover, the pairwise similarity approach often fails when many structural models are of low quality and similar to each other. To address the gap, we developed a hybrid method (MULTICOM_qa) combining a pairwise similarity score (PSS) and an interface contact probability score (ICPS) based on the deep learning inter-chain contact prediction for estimating protein complex model accuracy. It blindly participated in the 15th Critical Assessment of Techniques for Protein Structure Prediction (CASP15) in 2022 and ranked first out of 24 predictors in estimating the global accuracy of assembly models. The average per-target correlation coefficient between the model quality scores predicted by MULTICOM_qa and the true quality scores of the models of CASP15 assembly targets is 0.66. The average per-target ranking loss in using the predicted quality scores to rank the models is 0.14. It was able to select good models for most targets. Moreover, several key factors (i.e., target difficulty, model sampling difficulty, skewness of model quality, and similarity between good/bad models) for EMA are identified and analayzed. The results demonstrate that combining the multi-model method (PSS) with the complementary single-model method (ICPS) is a promising approach to EMA. The source code of MULTICOM_qa is available at https://github.com/BioinfoMachineLearning/MULTICOM_qa.

bioinformatics↗

bioRxiv2022

A deep reinforcement learning approach to reconstructing quaternary structures of protein dimers through self-learning

Predicted interchain residue-residue contacts can be used to build the quaternary structure of protein complexes from scratch. However, only a small number of methods have been developed to reconstruct protein quaternary structures using predicted interchain contacts. Here, we present an agent-based self-learning method based on deep reinforcement learning (DRLComplex) to build protein complex structures using interchain contacts as distance constraints. We rigorously tested the DRLComplex on two standard datasets of homodimeric and heterodimeric dimers (the CASP-CAPRI homodimer dataset and Std_32 heterodimer dataset) using both true and predicted contacts. Utilizing true contacts as input, the DRLComplex achieved a high average TM-score of 0.9895 and 0.9881 and a low average interface RMSD (I_RMSD) of 0.2197 and 0.92 on the two datasets, respectively. When predicted contacts are used, the method achieves the TM-score of 0.73 and 0.76 for homodimers and heterodimers respectively. The accuracy of reconstructed quaternary structures depends on the accuracy of contact predictions. Compared with other optimization methods of reconstructing quaternary structures from interchain contacts, DRLComplex performs similarly to an advanced gradient descent method and better than a Markov Chain Monte Carlo simulation method and a simulated annealing-based method. The source code of DRLComplex is available at: https://github.com/jianlin-cheng/DRLComplex

bioinformatics↗

bioRxiv2021

A deep dilated convolutional residual network for predicting interchain contacts of protein homodimers

MotivationDeep learning has revolutionized protein tertiary structure prediction recently. The cutting-edge deep learning methods such as AlphaFold can predict high-accuracy tertiary structures for most individual protein chains. However, the accuracy of predicting quaternary structures of protein complexes consisting of multiple chains is still relatively low due to lack of advanced deep learning methods in the field. Because interchain residue-residue contacts can be used as distance restraints to guide quaternary structure modeling, here we develop a deep dilated convolutional residual network method (DRCon) to predict interchain residue-residue contacts in homodimers from residue-residue co-evolutionary signals derived from multiple sequence alignments of monomers, intrachain residue-residue contacts of monomers extracted from true/predicted tertiary structures or predicted by deep learning, and other sequence and structural features. ResultsTested on three homodimer test datasets (Homo_std dataset, DeepHomo dataset, and CASP14-CAPRI dataset), the precision of DRCon for top L/5 interchain contact predictions (L: length of monomer in a homodimer) is 43.46%, 47.15%, and 24.81% respectively, which is substantially better than two existing deep learning interchain contact prediction methods. Moreover, our experiments demonstrate that using predicted tertiary structure or intrachain contacts of monomers in the unbound state as input, DRCon still performs reasonably well, even though its accuracy is lower than when true tertiary structures in the bound state are used as input. Finally, our case study shows that good interchain contact predictions can be used to build high-accuracy quaternary structure models of homodimers. AvailabilityThe source code of DRCon is available at https://github.com/jianlin-cheng/DRCon. Contactchengji@missouri.edu

bioinformatics↗

bioRxiv2021

Distance-based Reconstruction of Protein Quaternary Structures from Inter-Chain Contacts

Predicting the quaternary structure of a protein complex is an important and challenging problem. Inter-chain residue-residue contact prediction can provide useful information to guide the ab initio reconstruction of quaternary structures of protein complexes. However, few methods have been developed to build quaternary structures from predicted inter-chain contacts. Here, we introduce a new gradient descent optimization algorithm (GD) to build quaternary structures of protein dimers utilizing inter-chain contacts as distance restraints. We evaluate GD on several datasets of homodimers and heterodimers using true or predicted contacts. GD consistently performs better than a simulated annealing method and a Markov Chain Monte Carlo simulation method. Using true inter-chain contacts as input, GD can reconstruct high-quality structural models for homodimers and heterodimers with average TM-score ranging from 0.92 to 0.99 and average interface root mean square distance (I-RMSD) from 0.72 [A] to 1.64 [A]. On a dataset of 115 homodimers, using predicted inter-chain contacts as input, the average TM-score of the structural models built by GD is 0.76. For 46% of the homodimers, high-quality structural models with TM-score >= 0.9 are reconstructed from predicted contacts. There is a strong correlation between the quality of the reconstructed models and the precision and recall of predicted contacts. If the precision or recall of predicted contacts is >20%, GD can reconstruct good models for most homodimers, indicating only a moderate precision or recall of inter-chain contact prediction is needed to build good structural models for most homodimers. Moreover, the accuracy of reconstructed models positively correlates with the contact density in dimers and depends on the initial model and the probability threshold of selecting predicted contacts for the distance-based structure optimization.

bioinformatics↗

bioRxiv2020

Predicting interchain contacts for homodimeric and homomultimeric protein complexes using multiple sequence alignments of monomers and deep learning

Deep learning methods that achieved great success in predicting intrachain residue-residue contacts have been applied to predict interchain contacts between proteins. However, these methods require multiple sequence alignments (MSAs) of a pair of interacting proteins (dimers) as input, which are often difficult to obtain because there are not many known protein complexes available to generate MSAs of sufficient depth for a pair of proteins. In recognizing that multiple sequence alignments of a monomer that forms homomultimers contain the co-evolutionary signals of both intrachain and interchain residue pairs in contact, we applied DNCON2 (a deep learning-based protein intrachain residue-residue contact predictor) to predict both intrachain and interchain contacts for homomultimers using multiple sequence alignment (MSA) and other co-evolutionary features of a single monomer followed by discrimination of interchain and intrachain contacts according to the tertiary structure of the monomer. Allowing true-positive predictions within two residue shifts, the best average precision was obtained for the Top-L/10 predictions of DNCON2: 22.9% for homodimers, and 17.0% for higher order homomultimers. In some instances, especially where interchain contact densities are high, the approach predicted interchain contacts with 100% precision. We show that the predicted contacts can be used to accurately construct the structure of some complexes. Our experiment demonstrates that monomeric multiple sequence alignments can be used with deep learning to predict interchain contacts of homomeric proteins.

bioinformatics↗