Skip to main content
. 2022 Nov 28;32(9):1429–1438. doi: 10.1093/hmg/ddac289

Figure 3.

Figure 3

Classification of samples using SVM machine learning models based on the KBG DNAm signature. Sample groups were scored using the KBGS SVM model. KBGS validation subject (n = 7) classified as KBG-like demonstrating 100% sensitivity of the model. Whereas validation control subjects (n = 150) classified with low probabilities demonstrating 100% specificity of the model. ANKRD11 missense variants (n = 4), two of which belong to a child–parent duo and have KBG syndrome-like probabilities, whereas the remaining two had control-like probabilities.