The HNSCC metastasis signature outcome shows tumor cell percentage bias due to skewed distribution of signature components. (a) Metastatic signature profiles of seven analyzed primary HNSCCs based on: complete tumor sections and the originally identified 102-signature genes  (original); complete sections and the set of 685 metastasis associated predictive genes (complete); and the 685-gene set and synthetic samples in which the original tumor-stroma proportion was retained (lcm). Blue indicates a non-metastatic (N0) profile, and yellow indicates a metastatic (N+) profile. (b) Metastatic signature profiles of synthetic samples from 7 primary tumors that retained the original tumor percentage (lcm) or contained 0%, 25%, 50%, 75% or 100% tumor cells, respectively. Profiles are based on the predictive 685 gene set; colors are as in (a). (c) The set of 685 predictive genes are ordered according to the correlation of their expression level with the 35 analyzed tumor percentages. Colors are based on a direct microarray comparison of tumor cells and tumor stroma, which confirms that negatively correlated (<-0.50) genes are mainly expressed in the stroma and positively correlated gene (>0.50) are tumor cell associated. Green indicates higher expression in tumor stroma compared to tumor cells and red indicates higher expression in tumor cells than in tumor stroma. Which of the 685 signature genes are distributed over which different components is described in detail in Additional data file 1. (d) Tumor percentage correlation and signature association (N0 or N+) of the predictive genes. Tumor percentage correlative groups as shown in (c). Blue indicates genes that are associated with the N0 signature profile, and yellow those associated with an N+ profile. Stromal genes are mostly N+ associated, that is, with higher expression in N+ primary tumor sections, while N0 profile related predictive genes are more commonly expressed in tumor cells, that is, down-regulated in N+ primary tumors. (e) As (b), but for the tumor and stromal specific predictive genes (259 genes). (f) As (b), but for the non-specific predictive genes that are similarly expressed between tumor cells and tumor stroma (tumor percentage correlation between -0.50 and 0.50).