Table 3. Variation in gene expression levels within and between laboratories testing the same subtissue.
| Tissue category | Subtissue | Lab ID | Sample size | Within-lab correlation (r) | Between-lab correlation (r) |
|---|---|---|---|---|---|
| Brain+nerve | Caudate nucleus | E-AFMX-6 | 29 | 0.715 | 0.732 |
| GSE3790 | 1 | NA | |||
| Heart+muscle | Heart | GSE2240 | 23 | 0.722 | 0.880 |
| GSE974 | 13 | 0.606 | |||
| Heart+muscle | Smooth muscle | E-MEXP-569 | 8 | 0.749 | 0.570 |
| E-MEXP-66 | 1 | NA | |||
| Endocrine organs | Thymocyte | E-MEXP-337 | 13 | 0.623 | 0.853 |
| GSE1460 | 1 | NA | |||
| Brain+nerve | Brain | GSE5392 | 23 | 0.582 | 0.662 |
| E-LGCL-5 | 15 | 0.679 | 0.301 | ||
| E-TABM-145 | 1 | NA | 0.598 | ||
| Heart+muscle | Skeletal muscle | GSE3307 | 9 | 0.797 | 0.822 |
| GSE6011 | 7 | 0.803 | 0.793 | ||
| GSE1786 | 1 | NA | 0.449 |
|
Expression=subtissue+gene+lab(subtissue)+subtissue × gene+lab(subtissue) × gene | ||||
|---|---|---|---|---|
| Effect | d.f. | SS | F-value | P-value |
| Subtissue | 5 | 155.5 | 6.97 | 0.003 |
| Gene | 22 | 1003.7 | 11.86 | <0.001 |
| Lab(subtissue) | 8 | 22.8 | 1.1 | 0.362 |
| Subtissue × gene | 110 | 2510.6 | 6.14 | <0.001 |
| Lab(subtissue) × gene | 176 | 554.2 | 4.18 | <0.001 |
| Error | 3303 | 2489.0 | ||
The subtissues and their corresponding laboratories (represented by an ArrayExpress experiment ID) are given along with the number of tissue samples that were tested in each lab. For each lab (with sample size >1) an average within-lab correlation was calculated using the expression values across all 23 genes in each sample. For each subtissue, the pairwise correlation between all labs is given (in the order lab 1 vs 2, 1 vs 3, and 2 vs 3) using the mean expression values across all 23 genes for each lab. The analysis of variance (ANOVA) model and results of the data analysis are shown at the bottom of the table.