High-throughput sequencing has had an enormous impact on small RNA research during the past decade. However, sequencing only offers a one-dimensional view of the transcriptome and is often highly biased. Additionally, the ‘sequence, map and annotate’ approach, used widely in small RNA research, can lead to flawed interpretations of the data, lacking biological plausibility, due in part to database issues. Even in the absence of technical biases, the loss of three-dimensional information is a major limitation to understanding RNA stability, turnover and function. For example, noncoding RNA-derived fragments seem to exist mainly as dimers, tetramers or as nicked forms of their parental RNAs, contrary to widespread assumptions. In this perspective, we will discuss main sources of bias during small RNA-sequencing, present several useful bias-reducing strategies and provide guidance on the interpretation of small RNA-sequencing results, with emphasis on RNA fragmentomics. As sequencing offers a one-dimensional projection of a four-dimensional reality, prior structure-level knowledge is often needed to make sense of the data. Consequently, while less-biased sequencing methods are welcomed, integration of orthologous experimental techniques is also strongly recommended.

过去十年，高通量测序对小 RNA 研究产生了巨大影响。然而，测序仅提供转录组的一维视图，并且通常存在很大偏差。此外，在小RNA研究中广泛使用的“序列、图谱和注释”方法可能会导致对数据的错误解释，缺乏生物学合理性，部分原因是数据库问题。即使不存在技术偏差，三维信息的丢失也是理解 RNA 稳定性、周转和功能的主要限制。例如，非编码 RNA 衍生片段似乎主要以二聚体、四聚体或其亲代 RNA 的切口形式存在，这与普遍的假设相反。从这个角度来看，我们将讨论小 RNA 测序过程中偏差的主要来源，提出几种有用的减少偏差策略，并为小 RNA 测序结果的解释提供指导，重点是 RNA 片段组学。由于测序提供了四维现实的一维投影，因此通常需要先验的结构级知识来理解数据。因此，虽然偏差较小的测序方法受到欢迎，但也强烈建议整合直系同源实验技术。