I frequently encounter this issue of deflated lambda values (eg. 0.8 - 0.99) but the qqplot looks inflated, when running the qq function in a gwas study.
Why does this happen, is this due to technical error or with the study sample.?
Deflation shows that the study is underpowered so does the sample size has on effect on this issue?