Chromatin States Accurately Classify Cell Differentiation Stages
Figure 1
Overview of the data analysis strategy.
ChIPseq data of 5 histone modifications in 27 human cell lines were obtained from the public domains and analyzed independently using four different representations (bin, gene score, and chromatin state level). For each representation, a support vector machine (SVM) model was used to classify cell differentiation status from histone modification data. “Hotspot” genes or bins were detected by ANOVA and further investigated by functional genomic tools. The SVM model obtained from normal cell lines was used to classify the differentiation status of cancer cell lines.