Biology subjects

Xiang, J. S.

Publications and source records attributed to Xiang, J. S..

5 recordsLinked to original sources

bioRxiv2026

RNASeek: A Cross-Phyla Generative Foundation Model for Multipurpose RNA Modeling and Reinforcement Learning-Based Design

RNA plays central roles in regulating information flow and provides a versatile substrate for engineering biological functions. While large language models (LLMs) have transformed natural language processing and protein design, a general framework connecting RNA foundation models to functional sequence design remains limited. Here, we present RNASeek, a 1.6-billion-parameter generative foundation model built on a DeepSeek architecture and trained on a cross-phyla transcriptomic corpus for RNA sequence representation and generation. Natural-language tokens enable flexible conditional prediction and sequence design using a unified backbone. RNASeek captures species-specific transcript features and intron-exon boundaries in a zero-shot setting. We then fine-tune RNASeek to predict ribozyme self-cleavage activity and viral mRNA stability, revealing interpretable sequence features associated with function, including ribozyme loop flexibility and stem stability, as well as AU-rich motifs associated with mRNA stability. We use these functional predictors as reward models and apply Group Relative Policy Optimization (GRPO) to update the generation policy of RNASeek toward sequences with desired properties. GRPO-guided generation produces faster-cleaving ribozymes and stability-enhancing 3' UTRs while satisfying user-specified IUPAC constraints. Experimentally validated RNASeek-generated ribozymes achieve wild-type levels of activity, while RNASeek-generated 3' UTR sequences exceed the performance of the training data and benchmarked AI-generated 3' UTRs. Together, RNASeek establishes a unified pretrain-predict-optimize framework that connects learned RNA function to controllable de novo sequence design and provides a general strategy for engineering regulatory RNAs with desired properties.

bioengineering↗

bioRxiv2025

A conserved long-range RNA interaction in SARS-CoV-2 recruits ADAR1 to enhance virus proliferation

Long-range RNA-RNA pairing impacts the genome structure and function of SARS-CoV-2 variants. To understand the structure and function relationships of different SARS-CoV-2 variants that have emerged during the COVID-19 pandemic, we performed high-throughput structure probing and modelling of the genomic structures of the wildtype (WT), Alpha, Beta, Delta and Omicron variants of SARS-CoV-2. We observed that genomes of SARS-CoV-2 variants are generally structurally conserved, and that single-nucleotide variations and interactions with RNA binding proteins can impact RNA structures across the viruses. Importantly, using proximity ligation sequencing, we identified many conserved ultra-long-range RNA-RNA interactions, including one that spans more than 17 kb in both the WT virus and the Omicron variant. We showed that mutations that disrupt this 17 kb long-range interaction reduce viral fitness at later stages of its infection cycle, while compensatory mutations partially restore virus fitness. Additionally, we showed that this ultra-long-range RNA-RNA interaction structure binds directly to ADAR1 to alter the RNA editing levels on the viral genome. These studies deepen our understanding of RNA structures in the SARS-CoV-2 genome and their ability to interact with host factors to facilitate virus infectivity.

genomics↗

bioRxiv2025

Genome-Wide Interrogation of SARS-CoV-2 RNA-Protein Interactions Uncovers Hidden Regulatory Sites

The global impact of the COVID-19 pandemic underscores the critical need for a comprehensive understanding of SARS-CoV-2 replication mechanisms. While the central roles of the RNA dependent RNA polymerase (NSP12), primase protein (NSP8), and nucleocapsid protein (N) in the virus life cycle are extensively studied, the precise nature of their interactions with the full-length viral RNA genome remain incompletely characterized. In this study, we sought to address this knowledge gap by employing enhanced crosslinking and immunoprecipitation (eCLIP) to map the binding sites of NSP8, NSP12, and N proteins across the SARS-CoV-2 genome at early stages of viral RNA and protein synthesis and late stages of virion assembly. Our findings revealed interactions of NSP8 and NSP12 to the 5 and 3 untranslated regions (UTRs) of both positive and negative sense RNA, regions known to regulate viral replication, transcription, and translation. We identified a surprising and essential NSP12 binding site within the RNA sequence encoding the conserved Y1 domain of NSP3, which regulates RNA abundance upstream of the site. Additionally, we found that N protein interacts with the 5 UTR and influences translation efficiency. Finally, we report a novel regulatory function of N protein in modulating ribosomal frameshifting proximal to the frameshift element, a crucial process for maintaining viral protein stoichiometry. Our results provide a detailed molecular map of SARS-CoV-2 protein-RNA interactions, revealing potential therapeutic targets for attenuating viral fitness and informing the development of next-generation antiviral strategies.

molecular biology↗

bioRxiv2022

HydRA: Deep-learning models for predicting RNA-binding capacity from protein interaction association context and protein sequence

RNA-binding proteins (RBPs) control RNA metabolism to orchestrate gene expression, and dysfunctional RBPs underlie many human diseases. Proteome-wide discovery efforts predict thousands of novel RBPs, many of which lack canonical RNA-binding domains. Here, we present a hybrid ensemble RBP classifier (HydRA) that leverages information from both intermolecular protein interactions and internal protein sequence patterns to predict RNA-binding capacity with unparalleled specificity and sensitivity using support vector machine, convolutional neural networks and transformer-based protein language models. HydRA enables Occlusion Mapping to robustly detect known RNA-binding domains and to predict hundreds of uncharacterized RNA-binding domains. Enhanced CLIP validation for a diverse collection of RBP candidates reveals genome-wide targets and confirms RNA-binding activity for HydRA-predicted domains. The HydRA computational framework accelerates construction of a comprehensive RBP catalogue and expands the set of known RNA-binding protein domains. HighlightsO_LIHydRA combines protein-protein interaction and amino acid sequence information to predict RNA binding activity for 1,487 candidate genes. C_LIO_LIHydRA predicts RNA binding with higher specificity and sensitivity than current approaches, notably for RBPs without well-defined RNA-binding domains. C_LIO_LIOcclusion Mapping with HydRA enables RNA-binding domain discovery. C_LIO_LIEnhanced CLIP confirms HydRA RBP predictions with RNA-binding domain resolution. C_LI

bioinformatics↗

bioRxiv2022

Discovery and functional interrogation of SARS-CoV-2 protein-RNA interactions

The COVID-19 pandemic is caused by severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2). The betacoronvirus has a positive sense RNA genome which encodes for several RNA binding proteins. Here, we use enhanced crosslinking and immunoprecipitation to investigate SARS-CoV-2 protein interactions with viral and host RNAs in authentic virus-infected cells. SARS-CoV-2 proteins, NSP8, NSP12, and nucleocapsid display distinct preferences to specific regions in the RNA viral genome, providing evidence for their shared and separate roles in replication, transcription, and viral packaging. SARS-CoV-2 proteins expressed in human lung epithelial cells bind to 4773 unique host coding RNAs. Nine SARS-CoV-2 proteins upregulate target gene expression, including NSP12 and ORF9c, whose RNA substrates are associated with pathways in protein N-linked glycosylation ER processing and mitochondrial processes. Furthermore, siRNA knockdown of host genes targeted by viral proteins in human lung organoid cells identify potential antiviral host targets across different SARS-CoV-2 variants. Conversely, NSP9 inhibits host gene expression by blocking mRNA export and dampens cytokine productions, including interleukin-1/{beta}. Our viral protein-RNA interactome provides a catalog of potential therapeutic targets and offers insight into the etiology of COVID-19 as a safeguard against future pandemics.

molecular biology↗