Table 1. Analysis of random bias in TCGA datasets.
| Dataset | Signif % | P-value | PCNA % | PCNA p-val |
|---|---|---|---|---|
| ACC | 80 | 0 | 45 | 2.7e-122 |
| BLCA | 55 | 2e-187 | 51 | 1e-157 |
| BRCA | 21 | 1.3e-27 | 14 | 1.1e-12 |
| CESC | 6 | 0.38 | 7 | 0.072 |
| CHOL | 1 | 6.7e-10 | 1 | 2.2e-09 |
| COAD | 3 | 0.0063 | 1 | 5.7e-07 |
| COADREAD | 5 | 0.64 | 2 | 0.00012 |
| DLBC | 5 | 0.8 | 3 | 0.033 |
| ESCA | 0 | 1.1e-12 | 0 | 8.9e-12 |
| GBM | 7 | 0.075 | 7 | 0.11 |
| GBMLGG | 99 | 0 | 88 | 0 |
| HNSC | 25 | 8.5e-38 | 28 | 4.9e-48 |
| KICH | 8 | 0.0095 | 6 | 0.4 |
| KIPAN | 64 | 1.9e-279 | 24 | 1.3e-35 |
| KIRC | 82 | 0 | 68 | 0 |
| KIRP | 63 | 3.4e-260 | 10 | 4.1e-05 |
| LAML | 0 | 4.9e-11 | 0 | 1.1e-10 |
| LGG | 80 | 0 | 66 | 5.9e-302 |
| LIHC | 34 | 5.2e-70 | 4 | 0.35 |
| LUAD | 45 | 8.8e-118 | 19 | 3.8e-22 |
| LUSC | 20 | 1.9e-25 | 12 | 5.5e-09 |
| MESO | 53 | 7.7e-172 | 20 | 1.1e-25 |
| OV | 4 | 0.12 | 4 | 0.3 |
| PAAD | 43 | 2.1e-107 | 6 | 0.37 |
| PCPG | 4 | 0.16 | 4 | 0.17 |
| PRAD | 1 | 4.3e-06 | 1 | 3.9e-08 |
| READ | 2 | 0.00012 | 2 | 0.00061 |
| SKCM | 3 | 0.016 | 4 | 0.54 |
| TGCT | 0 | 2e-12 | 0 | 4e-13 |
| THCA | 3 | 0.0049 | 4 | 0.23 |
| THYM | 18 | 6.2e-21 | 20 | 2.7e-25 |
| UCEC | 62 | 1.1e-250 | 46 | 2.8e-127 |
| UCS | 2 | 1.7e-05 | 1 | 7.9e-06 |
| UVM | 38 | 5.3e-84 | 21 | 4e-28 |
In this table ‘Dataset’ indicates the abbreviation of the dataset as defined by the TCGA consortium; ‘Signif %’ is the proportion of significant random sets of size 64; ‘P-value’ is the significance by a proportion test for obtaining the proportion of significant %; ‘PCNA %’ is the proportion of significant random sets of size 64 after adjusting for the PCNA signature; and ‘PCNA p-val’ is the significance of the value in PCNA %. The analysis was performed with 5000 random sets.