Figure 2

Survival probability of the 43 samples. A) Prediction was based on the ${\widehat{v}}_1$ model using 30 transcripts with θ = 1.7 (${\widehat{v}}_1^{30}$). Dotted and solid curves represent samples with ${\widehat{v}}_1^{30}$ higher and lower than median ${\widehat{v}}_1$, respectively. Log-rank-p-value= 0.000127. B) The curves represent the predicted survival for patients obtained using different values of v1 (estimated using the first 1500 genes), i.e., the 10^th (0.044), 25^th (0.215), 50^th (0.280), 75^th (0.428), and 90^th (0.563) percentiles of v1.