dearseq: a variance component score test for RNA-Seq differential analysis that effectively controls the false discovery rate

Marine Gauthier (1, 2, 3, 4) , Denis Agniel (5, 6) , Rodolphe Thiébaut (1, 2, 3, 7, 4) , Boris Hejblum (1, 2, 3, 4)

1 SISTM - Statistics In System biology and Translational Medicine
2 BPH - Bordeaux population health
3 VRI - Vaccine Research Institute [Créteil, France]
4 UB - Université de Bordeaux
5 HMS - Harvard Medical School [Boston]
6 Rand Corporation
7 CHU Bordeaux

RNA-seq studies are growing in size and popularity. We provide evidence that the most commonly used methods for differential expression analysis (DEA) may yield too many false positive results in some situations. We present dearseq, a new method for DEA which controls the FDR without making any assumption about the true distribution of RNA-seq data. We show that dearseq controls the FDR while maintaining strong statistical power compared to the most popular methods. We demonstrate this behavior with mathematical proofs, simulations, and a real data set from a study of Tuberculosis, where our method produces fewer apparent false positives.

Mots clés: Differential expression; False Discovery Rate; RNA-seq; Score test; Tuberculosis; Type-I error; Variance component test

Domaines: Statistiques [math.ST]; Applications [stat.AP]; Méthodologie [stat.ME]

Marine Gauthier, Denis Agniel, Rodolphe Thiébaut, Boris Hejblum. dearseq: a variance component score test for RNA-Seq differential analysis that effectively controls the false discovery rate. NAR Genomics and Bioinformatics, 2020, 2 (4), pp.lqaa093. ⟨10.1093/nargab/lqaa093⟩. ⟨hal-02138664⟩