Benchmarking RNA-seq differential expression analysis methods using spike-in and simulation data

Abstract

Benchmarking RNA-seq differential expression analysis methods using spike-in and simulated RNA-seq data has often yielded inconsistent results. The spike-in data, which were generated from the same bulk RNA sample, only represent technical variability, making the test results less reliable. We compared the performance of 12 differential expression analysis methods for RNA-seq data, including recent variants in widely used software packages, using both RNA spike-in and simulation data for negative binomial (NB) model. Performance of edgeR, DESeq2, and ROTS was particularly different between the two benchmark tests. Then, each method was tested under most extensive simulation conditions especially demonstrating the large impacts of proportion, dispersion, and balance of differentially expressed (DE) genes. DESeq2, a robust version of edgeR (edgeR.rb), voom with TMM normalization (voom.tmm) and sample weights (voom.sw) showed an overall good performance regardless of presence of outliers and proportion of DE genes. The performance of RNA-seq DE gene analysis methods substantially depended on the benchmark used. Based on the simulation results, suitable methods were suggested under various test conditions.

WBP2 promotes BTRC mRNA stability to drive migration and invasion in triple‐negative breast cancer via NF‐κB activation

“…2D and Table ). The fold change in 1.23 is well within the benchmark that is widely adopted in transcriptomic studies [26,44,45]. Enrichment analysis of selected DEGs revealed that WBP2 is involved in multiple signaling pathways (Table 1).…”

Powerful p-value combination methods to detect incomplete association

“…10%, 30%, or 60% genes were made differentially expressed (DE) with 1.3 or larger fold changes. See our paper 20 for more detailed method to simulate read count data. The count data were then voom-transformed for moderated t -test (DE analysis) 21 .…”

In silico identification of novel open reading frames in Plasmodium falciparum oocyte and salivary gland sporozoites using proteogenomics framework

“…For RNA-seq data, DESeq 2 [53] was chosen to perform differential expression analysis, which is an R package available within the Bioconductor project [54]. This is because DESeq 2 is widely used by the community, including Lindner et al that published the analysed dataset [37], and it has also been recently proved to have the best overall performance among 12 methods [55]. Since DESeq2 requires read counts as an input, while StringTie outputs coverage values for transcript abundance, these were first converted from coverage to counts for each transcript, using the formula reads_per_transcript = coverage * transcript_len/read_len with a python script (available at http://ccb.jhu.edu/softw are/strin gtie/dl/ prepD E.py).…”