摘要As a single-stranded ribonucleic acid (RNA) virus, the replication, transcription, and interactions with host cells of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) rely on a complex network of RNA-RNA interactions. Investigating local RNA-RNA interactions is crucial for elucidating how viruses regulate their own functions and respond to host immune responses. This study aims to explore the application of machine learning techniques in analyzing and predicting RNA-RNA interactions within the coronavirus genome. Using virion RNA in situ conformation sequencing technology(vRIC-Seq) data and advanced computational models, we evaluated potential interactions between viral RNA fragments. By employing a variety of traditional machine learning algorithms, including traditional One-hot coding, Word2Vec models, a number of different neural network architectures, and the RNAErnie language modeling framework, we achieved significant predictive accuracy in determining the presence or absence of interactions. Furthermore, this approach provides a novel framework for investigating RNA-RNA interactions in other viral systems, thereby opening new avenues for the development of targeted therapeutic strategies against viral infections. The integration of computational models substantially enhances our comprehension of complex biological processes and represents a promising trajectory for future virology research. The source codes and models are freely available at https://github.com/VV1025/RNA-language-models.

abstractsAs a single-stranded ribonucleic acid (RNA) virus, the replication, transcription, and interactions with host cells of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) rely on a complex network of RNA-RNA interactions. Investigating local RNA-RNA interactions is crucial for elucidating how viruses regulate their own functions and respond to host immune responses. This study aims to explore the application of machine learning techniques in analyzing and predicting RNA-RNA interactions within the coronavirus genome. Using virion RNA in situ conformation sequencing technology(vRIC-Seq) data and advanced computational models, we evaluated potential interactions between viral RNA fragments. By employing a variety of traditional machine learning algorithms, including traditional One-hot coding, Word2Vec models, a number of different neural network architectures, and the RNAErnie language modeling framework, we achieved significant predictive accuracy in determining the presence or absence of interactions. Furthermore, this approach provides a novel framework for investigating RNA-RNA interactions in other viral systems, thereby opening new avenues for the development of targeted therapeutic strategies against viral infections. The integration of computational models substantially enhances our comprehension of complex biological processes and represents a promising trajectory for future virology research. The source codes and models are freely available at https://github.com/VV1025/RNA-language-models.