Figure 1

Distinct gene expression patterns in HCC from young or elderly patients. (A) A multidimensional scaling (MDS) plot using the whole transcriptome illustrates the mRNA profiles of normal and cancerous tissues. (B) A Venn diagram summarizing genes differentially expressed between the HCC tumor tissues of young and elderly patients, or between tumor and non-tumor yHCC samples. (C) Advanced signature training using a machine learning strategy and an independent testing elderly HCC data set. When probe sets were ranked by signal-to-noise ratios (weights), the top 309 features was the largest panel to give the lowest error rate (i.e., a best classification effect; upper panel). (D) The discrimination ability of the 309-probeset signature on the validation data set. The prediction strength plot (upper) and the PCA plot (lower) show the discriminating power of the identified 309 probe sets in separating young and elderly patients in the validation cohort. (E) A heat map shows the 309 probes sets differentiating young and elderly HCCs in the training data set 1, as well as discriminating tumor and non-tumor tissues. Columns represent tumor samples; rows represent probe sets. Genes in red: increased; in blue: decreased. Genes underlined: discussed in the text.