Table 2. Result summary of the all datasets analysed.
First column shows the names of datasets analysed. Numbers in parentheses show the sample sizes. ‘Among all genes’ column refers to the analyses performed using all genes relevant to those analyses (subcolumns) without a significance cutoff. ‘Within significant CoV changes’: genes show significant CoV change with age with FDR-corrected p-value<0.1. In the ‘DiCo vs. tissue specificity (Di- as background)’ column, divergent genes in development (Di-) were chosen as background. ‘Co vs. expression change in native tissue association (Figure 4b)’ column refers to the analysis performed in Figure 4b for each dataset, and the results are presented in Figure 4c. The association tests were performed among convergent genes in ageing except in our dataset, which was performed with DiCo genes. Significant test results are indicated with italic fonts. Bold fonts show the results that support convergence or tissue-specific expression loss in ageing whether as a significant result or as a trend. Unsupportive test results and inapplicable tests are written in normal font.
| Among all genes | Within significant CoV changes | ||||||
|---|---|---|---|---|---|---|---|
| PCA change in Euclidean distance | Mean CoV change | Median CoV change | Pairwise tissue correlations | DiCo vs. tissue specificity(Di- as background) | Co vs. expression change in native tissue association (Figure 4b) | Co vs. Di proportions | |
| Izgi2022 |
ρ = −0.87,
p=0.0026 |
ρ = −0.5, p=0.2 | rho = −0.48, P = 0.23 | 4/6 positive, none significant* | OR = 1.56, p=1.3 × 10–18 | OR = 74.81, p=5.9 × 10–203 (among 1287 DiCo genes) | 68% convergence(among 62 significant genes*) |
| Jonker 2013five tissues, two different than ours (n = 18) |
ρ=−0.57,
p=0.014 |
ρ=−0.48, p=0.044 | ρ = −0.03, p=0.91 | 7/10 positive, none significant* | Di- background missing | OR = 7.52, p=6.5 × 10 –109 (among 2967 convergent genes) | 66% convergence (among 1735 significant genes*) |
| Schaum 2020,same four tissues (n = 37) | ρ = 0.13, p=0.46 |
ρ = 0.25, p=0.14 | ρ = 0.13, p=0.43 | 4/6 positive, two significant* | OR = 1.33, p=1.07 × 10–8 | OR = 58.03, p=1.5 × 10–197 (among 2124 convergent genes) | 53% convergence (among 319 significant genes*) |
| Schaum 2020,eight tissues (n = 26) | ρ = 0.1, p=0.62 |
ρ = 0.16, p=0.43 | ρ = 0.04, p=0.86 | 16/28 positive, five significant* | Di- background missing | OR = 84.2, p=9.7 × 10–96 (among 2380 convergent genes) | 54% convergence (among 244 significant genes*) |
| GTEx,same four tissues |
ρ = −0.23, p=0.12 |
ρ = −0.12, p=0.42 | ρ = −0.18, p=0.23 | 5/6 positive, none significant* | Di- background missing | OR = 7.21, p=7 × 10–87 (among 2407 convergent genes) | (no significant CoV changes) |
| GTEx,10 tissues |
ρ = −0.26, p=0.13 |
ρ = −0.14, p=0.44 | ρ = −0.3, p=0.08 | 29/45 positive, none significant* | Di- background missing | OR = 13.01, p=5.7 × 10–114 (among 2195 convergent genes) | (all three significant genes were convergent) |
ρ = Spearman’s correlation coefficient; OR = odds ratio; FDR = false discovery rate; CoV = coefficient of variation; DiCo = divergence-convergence; PCA = principal components analysis.
* FDR-corrected p-value<0.1.