Table 7.
Analysis of variance performed on the average concordance rate and allelic R 2 of the animals in the imputation population from each scenario for imputation from low density panels to the 777K SNP panel 1,2,3
| Concordance rate | Allelic R 2 | ||||
|---|---|---|---|---|---|
| Source | Mean | Scheffé test 4 | Source | Mean | Scheffé test 4 |
| Algorithm 5 (P-value < 0.0001) | Algorithm 5 (P-value?<?0.0001) | ||||
| FImpute | 1.291 | a | FImpute | 1.203 | a |
| Beagle | 1.244 | b | Beagle | 1.145 | b |
| Panel 6 (P-value < 0.0001) | Panel 6 (P-value?<?0.0001) | ||||
| 90tK | 1.351 | a | 90tK | 1.286 | a |
| 90iK | 1.343 | b | 90iK | 1.275 | b |
| 50K | 1.295 | c | 50K | 1.210 | c |
| 15K | 1.273 | d | 15K | 1.181 | d |
| 20K | 1.247 | e | 20K | 1.146 | e |
| 8K | 1.245 | e | 8K | 1.144 | e |
| 6K | 1.239 | e | 6K | 1.135 | e |
| 3K | 1.150 | f | 3K | 1.013 | f |
| Scenario 7 (P-value 0.0258) | Scenario (P-value 0.0346) | ||||
| NE-NP | 1.269 | a | NE-NP | 1.175 | a |
| NE-P | 1.267 | b | NE-P | 1.172 | b |
| Algorithm*panel (P-value =0.0052) | Algorithm*panel (P-value =0.0107) | ||||
| FImpute - 90tK | 1.370 | a | FImpute - 90tK | 1.309 | a |
| FImpute - 90iK | 1.364 | a | FImpute - 90iK | 1.301 | a |
| Beagle - 90tK | 1.331 | b | Beagle - 90tK | 1.262 | b |
| FImpute - 50K | 1.322 | b | Beagle - 90iK | 1.249 | b |
| Beagle - 90iK | 1.322 | b | FImpute - 50K | 1.244 | b |
| FImpute - 15K | 1.300 | c | FImpute - 15K | 1.215 | c |
| FImpute - 20K | 1.271 | d | Beagle - 50K | 1.176 | d |
| FImpute - 8K | 1.269 | d | FImpute - 20K | 1.176 | d |
| Beagle - 50K | 1.269 | d | FImpute - 8K | 1.174 | d |
| FImpute - 6K | 1.262 | d | FImpute - 6K | 1.165 | d |
| Beagle - 15K | 1.245 | e | Beagle - 15K | 1.146 | e |
| Beagle - 20K | 1.222 | f | Beagle - 20K | 1.115 | f |
| Beagle - 8K | 1.221 | f | Beagle - 8K | 1.114 | f |
| Beagle - 6K | 1.215 | f | Beagle - 6K | 1.106 | f |
| FImpute - 3K | 1.169 | g | FImpute - 3K | 1.039 | g |
| Beagle - 3K | 1.130 | h | Beagle - 3K | 0.988 | h |
1Concordance rate and allelic R2 were arcsine square root transformed for the analyses;
2Interaction effects between Algorithm*Scenario and Panel*Scenario were not statistically significant (P?>?0.05);
33K, 6K, 8K, 15K and 20K are low-density panels were imputed in two steps (firstly they were imputed to the 50K and then to the 777K SNP panel);
4Different letters within a group means that there is a statistical difference between two means (P?<?0.05);
5Algorithm used was either FImpute v.2.2 [11] or Beagle v.3.3 [8];
63K, 6K, 8K, 15K, 20K, 50K, 90iK and 90tK are low-density panels;
7Scenarios for imputation to the 777K SNP panel. (NE-P) - using Nellore genotypes in the reference population and considering pedigree information; (NE-NP) - using Nellore genotypes in the reference population and not using pedigree information.