blitzGSEA: efficient computation of gene set enrichment analysis through gamma distribution approximation
blitzGSEA: efficient computation of gene set enrichment analysis through gamma distribution approximation
复制标题
DOI：
10.1093/bioinformatics/btac076
复制
发表时间：
2022-02-10
期刊：
影响因子：
5.8
通讯作者：
Ma'ayan, Avi
中科院分区：
文献类型：
--
作者：
Lachmann, Alexander;Xie, Zhuorui;Ma'ayan, Avi
从差异基因表达中识别途径和生物学过程对于解释转录组学分析收集的数据至关重要。基因集富集分析（GSEA）是计算具有差异表达签名的注释基因集的相关性的显著性的最常用的算法。为了计算显着性，GSEA实现置换测试，这是缓慢和不准确的比较许多差异表达签名的数千个注释gene sets.Results在这里，我们提出了blitzGSEA，一种算法，是基于相同的运行和统计GSEA，但不是执行置换，blitzGSEA近似的富集分数概率的基础上伽玛分布。与之前的GSEA实现相比，blitzGSEA在性能上实现了显着改进，同时更准确地逼近小P值。
Motivation The identification of pathways and biological processes from differential gene expression is central for interpretation of data collected by transcriptomics assays. Gene set enrichment analysis (GSEA) is the most commonly used algorithm to calculate the significance of the relevancy of an annotated gene set with a differential expression signature. To compute significance, GSEA implements permutation tests which are slow and inaccurate for comparing many differential expression signatures to thousands of annotated gene sets.Results Here, we present blitzGSEA, an algorithm that is based on the same running sum statistic as GSEA, but instead of performing permutations, blitzGSEA approximates the enrichment score probabilities based on Gamma distributions. blitzGSEA achieves significant improvement in performance compared with prior GSEA implementations, while approximating small P-values more accurately.