Figure 4.

Combined variance and response-related filtering of the NCI-60 data. (a) The projection score for various choices of the inclusion criteria for the variance (θ, fraction of max variance) and the p-value from an F-test contrasting all nine cancer types (log₁₀(α), where α is the p-value threshold). The optimal projection score is obtained by combining the two filtering procedures. (b) The sample representation obtained by applying PCA to the most informative variable subset. In panel (b), in order to obtain a more easily interpretable plot, we joined the closest neighbors among the samples with line segments. The distance between two samples is defined by the Euclidean distance in the space spanned by all the remaining variables. The figure in (b) was generated using Qlucore Omics Explorer 2.2 (Qlucore AB, Lund, Sweden).