Entering edit mode
12.6 years ago
Psb
▴
30
Hi!! I am using EIGENSTRAT to detect if my samples contains clusters of sub-populations. The p-value along each eigenvector for population differences are insignificant, p values are more than 0.01.
eigenvector_1_Control_Case_ 0.107068
eigenvector_2_Control_Case_ 0.158401
eigenvector_3_Control_Case_ 0.619718
eigenvector_4_Control_Case_ 0.372473
eigenvector_5_Control_Case_ 0.740483
eigenvector_6_Control_Case_ 0.672963
eigenvector_7_Control_Case_ 0.91454
eigenvector_8_Control_Case_ 0.492866
eigenvector_9_Control_Case_ 0.39202
eigenvector_10_Control_Case_ 0.288796
Following are the values for co-rrelation between eigenvector and case-control status
Correlation between eigenvector 1 (of 10) and Case/Control status is 0.045
Correlation between eigenvector 2 (of 10) and Case/Control status is 0.040
Correlation between eigenvector 3 (of 10) and Case/Control status is 0.014
Correlation between eigenvector 4 (of 10) and Case/Control status is 0.025
Correlation between eigenvector 5 (of 10) and Case/Control status is 0.009
Correlation between eigenvector 6 (of 10) and Case/Control status is 0.012
Correlation between eigenvector 7 (of 10) and Case/Control status is -0.003
Correlation between eigenvector 8 (of 10) and Case/Control status is 0.019
Correlation between eigenvector 9 (of 10) and Case/Control status is 0.024
Correlation between eigenvector 10 (of 10) and Case/Control status is 0.030
Does this conclude that the spread of samples seen in the plot is not due to the population stratification?? Please help me understand this concept.