NBAMSeq

"NBAMSeq" introduces a flexible statistical model for high-throughput sequencing experiments tailored explicitly for differential expression analysis. Unlike existing models that assume a linear effect of covariates, NBAMSeq employs a generalized additive model, allowing for the detection of nonlinear effects in gene counts associated with relevant covariates. The model incorporates information sharing across genes in variance estimation, enhancing performance. Extensive simulations and case studies on RNA-Seq data demonstrate that NBAMSeq outperforms existing methods in detecting nonlinear effects while maintaining comparable performance in detecting linear effects.