Alternative polyadenylation (APA) generates multiple transcripts with alternative 3' untranslated regions (3' UTRs) by employing different polyadenylation sites (PASs), which influence subcellular localization of both RNA and protein, stability of mRNAs, and translation efficiency (Berkovits and Mayr, 2015; Tian and Manley, 2017). Intronic polyadenylation (IPA) could recognize premature termination signals and lead to changes in either the coding sequence or the 3′ UTR of the corresponding mRNA (Zhao et al., 2021b). While 3′ UTR APA is widespread and extensively exploited in diverse biological processes (Sandberg et al., 2008; Gruber and Zavolan, 2019; Mitschka and Mayr, 2022; Rossi and Kontarakis, 2022; Tang et al., 2022), IPA is less prevalent and their biological functions are not well understood. Existing research revealed the prominence of IPA sites in diverse biological processes and pathological conditions, particularly in cancer development (Tian et al., 2007; Dubbury et al., 2018; Lee et al., 2018; Singh et al., 2018). Recent studies also highlight the biological significance of IPA. As an example, CDK13 mutations impede an RNA surveillance mechanism to degrade abnormal RNAs, which causes accumulation and translation of aberrant prematurely terminated RNAs (ptRNAs) ending in introns (Insco et al., 2023). Mislocalization of nuclear TDP-43 triggers the utilization of STMN2 IPA sites, leading to the generation of a truncated RNA that encodes a non-functional fragment of stathmin-2 (Baughn et al., 2023).

Dysregulated IPA events during tumorigenesis have been reported to play important roles in cancer development. We have previously developed a bioinformatic algorithm called IPAFinder to perform de novo identification and quantification of IPA events from standard RNA-seq data (Zhao et al., 2021b). Applying IPAFinder to TCGA RNA-seq data, we found that somatic variants near splice site in ten cancer types could trigger abnormal IPA events, which serve as a novel potential mechanism of tumor suppressor genes inactivation (Zhao et al., 2021a). However, a comprehensive landscape of IPA events across all TCGA cancer types is still lacking, hampering the fully understanding of IPA regulation and its clinical utility. To address this knowledge and information gap, we developed a comprehensive database named IPACancer Atlas for users to search and explore the dynamically changed, clinically relevant, drug response and single nucleotide variant (SNV)-associated IPA events across all 33 TCGA cancer types.