Correlation between accuracy sensitivity specificity and positive predictive value parameters in detecting differentially expressed genes

RNA-seq high-throughput sequencing technology is rapidly becoming the standard method for measuring RNA expression levels. We previously evaluated the performance of different grape microarray design strategies based on custom microarray platforms assuming RNA-seq gene expression data as a reference. We subsequently evaluated sensitivity, specificity, accuracy and positive predictive value (PPV) parameters of these microarray design strategies in detecting significantly differential expressed genes (DEGs) in gene expression differential analysis. For this paper we investigated the relationship between these parameters applying several R software statistical tests and functions. This survey emphasizes a strong discrepancy between sensitivity and specificity parameters (p-value ≤ 0.001) evaluating the analyzed grape microarray design strategies performance in discriminating significantly differentially expressed genes. Furthermore, we demonstrated a substantial correlation and a lower variance difference between specificity, accuracy and PPV parameters estimating microarrays’ capacity to detect differentially expressed genes.