Table 1 The results of different distances with the row for DTW highlighted in bold being the best choice (the missed number of genes from other clusters are potentially significant genes)
From: Inferring circadian gene regulatory relationships from gene expression data with a hybrid framework
Distance | Rank in | In Cluster 6 | #missed | |
|---|---|---|---|---|
total | Rank | #significant | genes from | |
genes | Â | genes | other clusters | |
1. Manhattan | 1 | 1 | 823 | 187 |
2. Dissim | 1 | 1 | 780 | 173 |
3. Cort | 1 | 1 | 791 | 177 |
4. Euclidean | 2 | 2 | 739 | 108 |
5. Cor | 2 | 2 | 739 | 168 |
6. Fourier | 2 | 2 | 741 | 171 |
7. Ccor | 5 | 5 | 63 | 18 |
8. DTW | 7 | 7 | 253 | 51 |
9. Minkowski | 7 | 7 | 584 | 133 |
10. ACF | 30 | 21 | 35 | 15 |
11. TAM | 32 | 31 | 264 | 40 |
12. STS | 34 | 24 | 34 | 16 |
13. SPEC | 63 | 39 | 3 | 17 |
14. Periodogram | 64 | 31 | 9 | 3 |
15. Infinite norm | 97 | 96 | 313 | 174 |
16. INTPER | 121 | 72 | 3 | 6 |
17. Frechet | 166 | 156 | 278 | 135 |
18. NCD | 4219 | 391 | 3 | 16 |
19. CDM | 4952 | 396 | 3 | 65 |
20. PACF | 17049 | 955 | 3 | 123 |