Table 1. Number of SNPs significantly correlated with top low dimensional coordinates and PCs.
Population | CEU | YRI | ASI |
Individuals | 60 | 59 | 60 |
P-value for significance | 6.51E-09 | 4.75E-09 | 8.23E-09 |
All SNPs | 7,681,165 | 10,518,225 | 6,077,954 |
LLE Significant | 3,099,007 | 6,216,648 | 1,740,342 |
PCA Significant | 3,413,850 | 6,124,608 | 1,272,679 |
LLE (%) | 40.3% | 59.1% | 28.6% |
PCA (%) | 44.4% | 58.2% | 20.9% |
Common SNPs | 5,853,886 | 7,382,215 | 5,259,537 |
LLE Significant | 1,468,999 | 3,318,249 | 1,125,869 |
PCA Significant | 1,860,653 | 3,266,439 | 1,172,668 |
LLE (%) | 25.1% | 44.9% | 21.4% |
PCA (%) | 31.8% | 44.2% | 22.3% |
Rare SNPs | 1,827,279 | 3,136,010 | 818,417 |
LLE Significant | 1,630,008 | 2,898,399 | 614,473 |
PCA Significant | 1,553,197 | 2,858,169 | 100,011 |
LLE (%) | 89.2% | 92.4% | 75.1% |
PCA (%) | 85.0% | 91.1% | 12.2% |
Ancestry-informative markers in the LLE and PCA are identified by a simple regression where the low dimensional coordinates and principal component score are regressed on a genomic variant variable. SNPs with P-values less than threshold after Bonferroni correction for multiple tests are selected as LLE or PCA significantly correlated markers.