Skip to main content
Fig. 5 | Biology Direct

Fig. 5

From: Genes sharing the protein family domain decrease the performance of classification with RNA-seq genomic signatures

Fig. 5

Violinplot of misclassified samples for 4 classifiers based on DEGs with and without domains. From the differentially expressed genes with the significance level α=0.05 we choose two subsets: the first one was the genes that share one particular domain (with the biggest number of genes connected to this domain) and the second was the genes that share no domain. Validation was performed with 5 fold cross-validation. Percentages of misclassified samples are mostly lower for the cases where genes with no domains are taken into account

Back to article page