LDPAC is a method for detecting spurious associations using p-values. The strength of the method is its simplicity; it requires no other information but p-values. We provide both the Web server where user can upload p-values, and the Software binary. Our method uses internally stored HapMap Phase III data to retrieve LD information.
The intuition is that if the association is caused by a true genetic effect, neighboring markers in LD should show comparably significant p-values (Figure A). If not, it may be evidence of spurious association (Figure B). Previous studies have applied the same idea but only manually without a formal statistical framework. Our method LDPAC provides a systematic framework to quantitatively measure the evidence of spurious associations.
We take WTCCC CAD chr 9 data
where a confirmed association exists.
We put 5 spurious associations back into the data
which were originally excluded by QC in WTCCC study.
Figure shows that our method effectively detects 3
out of 5 spurious associations (red dots)
using only p-values.
The true association is correctly predicted as true (blue dots).
