Stochastic principles governing alternative splicing of RNA

The dominance of the major transcript isoform relative to other isoforms from the same gene generated by alternative splicing (AS) is essential to the maintenance of normal cellular physiology. However, the underlying principles that determine such dominance remain unknown. Here, we analyzed the physical AS process and found that it can be modeled by a stochastic minimization process, which causes the scaled expression levels of all transcript isoforms to follow the same Weibull extreme value distribution. Surprisingly, we also found a simple equation to describe the median frequency of transcript isoforms of different dominance. This two-parameter Weibull model provides the statistical distribution of all isoforms of all transcribed genes, and reveals that previously unexplained observations concerning relative isoform expression derive from these principles.

由可变剪接（alternative splicing，AS）产生的同一基因的各转录本亚型中，优势转录本亚型相较于其余亚型的表达优势，对维持正常细胞生理功能至关重要。然而，决定此类表达优势的底层原理仍未明晰。本研究对可变剪接的物理过程展开分析后发现，该过程可通过随机最小化过程建模，该过程使得所有转录本亚型的标准化表达水平均服从同一威布尔极值分布（Weibull extreme value distribution）。令人意外的是，本研究还找到了一个简洁方程，可用于描述不同优势程度的转录本亚型的中位频率。该双参数威布尔模型可给出所有转录基因的全部转录本亚型的统计分布，并揭示出此前关于转录本亚型相对表达的诸多无法解释的观测结果，实则均源于上述原理。