Benchmarking Methods for Detecting Present/Absent Genes Using RNA-seq Data

Accurately identifying whether genes are expressed or inactive in RNA-seq datasets remains challenging due to biological and technical noise. Traditional threshold-based approaches don't take into account noise differences between experiments and species.. Here, we benchmark advanced statistical and Bayesian methods—including Zigzag, RefBool scGenexpress,, BulkECexplorer, DAFS, and BgeeCall to robustly distinguish true gene expression from noise, highlighting performance and reliabilityy across datasets.