Skip to main content
Fig. 2 | BMC Bioinformatics

Fig. 2

From: PyGNA: a unified framework for geneset network analysis

Fig. 2

GNT benchmarking. a Performance of all GNT tests on the SBM networks. We show True Positive Rate (TPR) and False Positive Rate (FPR) (y-axis) of each GNT test (colors) for different values of \(\alpha\) (x-axis). As expected, as the value of \(\alpha\) increases, all tests improve their detection performance, with \(T_H\) and \(T_{ID}\) having consistently TPR \(>0.75\). Conversely, for FPR we do not see a strong effect as \(\alpha\) increases, with most tests having FPR \(\sim 5\%\). b Extended geneset high degree nodes (HDNs) networks used to quantify FPR. Genesets have been selected with increasing number of HDNs (x-axis) and random nodes to HDNs ratios (colors); for each analysis, we report the False Positive Rate (FPR). As the ratio between random and HDNs increases (\(\rho\)), we notice that \(T_{SP}\) has better performances. Interestingly, \(T_{ID}\) is the only one with FPR \(<5\%\) in all conditions

Back to article page