Table 2.
Multivariate linear models explaining the evolutionary distance between samples
| Predictor of evolutionary distance | Estimate | SE | t-Value | p |
|---|---|---|---|---|
| a) Single-crypt study (8 patients) | ||||
| (Intercept) | 14.06 | 12.27 | 1.15 | 0.5691 |
| Progressor status (yes) | 18.91 | 5.49 | 3.44 | 0.0017 |
| Time point (second) | 6.46 | 6.123 | 1.06 | 0.0537 |
| Maximum distance from GEJ (cm) | −1.48 | 0.52 | −2.84 | 0.0088 |
| Distance between biopsies (cm) | −1.03 | 0.93 | −1.11 | 0.2545 |
| b) Large cohort (197 patients) | ||||
| (Intercept) | 4.54 | 1.50 | 3.03 | 0.0025 |
| Progressor (yes) | 12.75 | 0.78 | 16.30 | <2e −16 |
| Time point (second) | −0.25 | 0.77 | −0.33 | 0.7430 |
| Maximum distance from GEJ (cm) | −0.50 | 0.15 | −3.21 | 0.0014 |
| Distance between biopsies (cm) | 0.57 | 0.20 | 2.81 | 0.0051 |
a) For the eight patients in our study, genetic distances between crypts are based on phylogenetic estimates of the number of genetic events that separate the samples. b) For the larger cohort of Barrett’s patients, the genetic distance between biopsies was calculated based on divergent copy number calls across 1 Mb windows as described in11. Values highlighted in bold pass the 0.05 threshold for statistical significance.