ERFMTDA: Predicting tsRNA–disease associations using an enhanced rotative factorization machine

Abstract Motivation tRNA-derived small RNAs (tsRNAs) have emerged as a novel class of regulatory molecules implicated in the pathogenesis of many human diseases, making them as promising biomarkers and therapeutic targets. However, existing computational methods for tsRNA–disease association prediction often overlook explicit biological attributes and complex feature interactions, limiting their predictive performance. Results We propose ERFMTDA, an enhanced rotative factorization machine framework for predicting potential tsRNA–disease associations. ERFMTDA explicitly models complex interactions among heterogeneous biological features while integrating latent structural representations derived from the global association matrix. In addition, a biologically informed negative sampling strategy based on motif-level sequence similarity is introduced to improve the reliability of negative samples. Extensive experiments demonstrate that ERFMTDA consistently outperforms eleven state-of-the-art methods. Case studies on diabetic retinopathy and hepatocellular carcinoma further confirm its ability to prioritize biologically meaningful tsRNA–disease associations. Availability and implementation The source codes and datasets of ERFMTDA are available at https://github.com/lanbiolab/ERFMTDA .