Accumulating studies demonstrated that the roles of lncRNAs for tumorigenesis were isoform-dependent and their aberrant splicing patterns in cancers contributed to function specificity. However, there is no existing database focusing on cancer-related alternative splicing of lncRNAs. Here, we developed a comprehensive database called LncAS2Cancer, which collected 5335 bulk RNA sequencing and 1826 single-cell RNA sequencing samples, covering over 30 cancer types. By applying six state-of-the-art splicing algorithms, 50 859 alternative splicing events for 8 splicing types were identified and deposited in the database. In addition, the database contained the following information: (i) splicing patterns of lncRNAs under seven different conditions, such as gene interference, which facilitated to infer potential regulators; (ii) annotation information derived from eight sources and manual curation, to understand the functional impact of affected sequences; (iii) survival analysis to explore potential biomarkers; as well as (iv) a suite of tools to browse, search, visualize and download interesting information. LncAS2Cancer could not only confirm the known cancer-associated lncRNA isoforms but also indicate novel ones. Using the data deposited in LncAS2Cancer, we compared gene model and transcript overlap between lncRNAs and protein-coding genes and discusses how these factors, along with sequencing depth, affected the interpretation of splicing signals. Based on recurrent signals and potential confounders, we proposed a reliable score to prioritize splicing events for further elucidation. Together, with the broad collection of lncRNA splicing patterns and annotation, LncAS2Cancer will provide important new insights into the diverse functional roles of lncRNA isoforms in human cancers. LncAS2Cancer is freely available at https://lncrna2as.cd120.com/.

中文翻译：

---

LncAS2Cancer：用于人类癌症中 lncRNA 选择性剪接的综合数据库。

越来越多的研究表明，lncRNA 在肿瘤发生中的作用是异构体依赖性的，它们在癌症中的异常剪接模式有助于功能特异性。然而，没有现有的数据库专注于 lncRNA 的癌症相关选择性剪接。在这里，我们开发了一个名为 LncAS2Cancer 的综合数据库，该数据库收集了 5335 个批量 RNA 测序和 1826 个单细胞 RNA 测序样本，涵盖了 30 多种癌症类型。通过应用六种最先进的剪接算法，8 种剪接类型的 50 859 个备选剪接事件被识别并存入数据库。此外，该数据库还包含以下信息：(i) lncRNAs 在七种不同条件下的剪接模式，例如基因干扰，有助于推断潜在的调节因子；(ii) 来自八个来源和人工管理的注释信息，以了解受影响序列的功能影响；(iii) 生存分析以探索潜在的生物标志物；以及 (iv) 一套用于浏览、搜索、可视化和下载有趣信息的工具。LncAS2Cancer 不仅可以确认已知的癌症相关 lncRNA 亚型，还可以指示新的亚型。使用保存在 LncAS2Cancer 中的数据，我们比较了 lncRNA 和蛋白质编码基因之间的基因模型和转录本重叠，并讨论了这些因素以及测序深度如何影响剪接信号的解释。基于反复出现的信号和潜在的混杂因素，我们提出了一个可靠的评分来确定剪接事件的优先级，以便进一步阐明。连同广泛的 lncRNA 剪接模式和注释集合，LncAS2Cancer 将为 lncRNA 亚型在人类癌症中的不同功能作用提供重要的新见解。LncAS2Cancer 可在 https://lncrna2as.cd120.com/ 免费获得。