DeepRetention: A Deep Learning Approach for Intron Retention Detection

• DOI: 10.26599/bdma.2022.9020023
• Publication Date: 2023-01-25T18:42:51Z
• Authors: Zhenpeng Wu; Jiantao Zheng; Jiashu Liu; Cuixiang Lin; Hong-Dong Li
• Keywords: RNA-Seq; Gold standard (test)

As the least understood mode of alternative splicing, Intron Retention (IR) is emerging as an interesting area and has attracted more attention in field gene regulation disease studies. Existing methods detect IR exclusively based on one or a few predefined metrics describing local summarized characteristics retained introns. These are not able to describe pattern sequencing depth intronic reads, which intuitive informative characteristic We hypothesize that incorporating distribution reads will improve accuracy detection. Here we present DeepRetention, novel approach for detection by modeling Due lack gold standard dataset IR, first compare DeepRetention with two state-of-the-art methods, i.e. iREAD IRFinder, simulated RNA-seq datasets The results show outperforms these methods. Next, performs well when it applied third-generation long-read data, while IRFinder applicable detecting from data. Further, IRs predicted biologically meaningful Alzheimer's Disease (AD) samples. differential found be significantly associated AD statistical evaluation AD-specific functional network. parent genes enriched AD-related functions. In summary, detects new angle view, providing valuable tool analysis.