Table 1.
Multivariate association analysis of fragment-length alleles within both VNTRs and the common 3975 bp deletion in the DOCK5 region
| Allele frequency |
β (SE) | Association with obesity |
||||
|---|---|---|---|---|---|---|
| Child | Adult | Child | Adult | Child | Adult | |
| 3975 bp deletion | 17%a | 18%a | −1.2 (0.5) | −0.5 (0.3) | 1.1 × 10−2 | 5.2 × 10−2 |
| VNTR-selected variables | ||||||
| VNTR A: 590–640 bp | 1.3% | 1.4% | 0.8 (0.3) | 1.1 (0.4) | 4.9 × 10−3 | |
| VNTR B: 944–1022 bp | 20% | 16% | −0.4 (0.1) | 0.1 (0.1) | 0.27 | |
| VNTR B: 1112–1127 bp | 1.9% | 1.2% | −1.5 (0.4) | 0.1 (0.4) | 0.68 | |
| VNTR B: 1073–1084 bp | 3.9% | 2.0% | −1.5 (0.3) | −0.6 (0.3) | 6.4 × 10−2 | |
| VNTR B: 1099–1103 bp | 1.6% | 1.2% | −0.9 (0.4) | 0.8 (0.4) | 8.3 × 10−2 | |
| VNTR A + B multivariate model | 8.9 × 10−8 | 3.1 × 10−3 | ||||
Associations between the 3,975 bp deletion and obesity were assessed under a recessive model in the child and adult obesity case-control samples. Multivariate models were trained in the child obesity case–control study and replicated in the adult obesity study. Models were built for VNTR A and subsequently VNTR B. Effect sizes and standard errors, as well as replication (adult) P-values were calculated using a logistic regression model. The VNTRA + VNTRB model significance is calculated using a likelihood ratio test between a model that includes all variables and a model including only the 3975 bp deletion and gender. The model significance presented for the child training data set is the empirical P-value (10 000 permutations).
aNote that the allele frequency presented in this table for the 3975 bp deletion in DOCK5 is that of the rarer undeleted allele at this locus.