I ran a few quick simulations with quartic trends and local linear regression, but used bw's at different multiples of the CCT bw. As expected, using uncorrected estimates, the MSE is minimized at the CCT bw, but the bias-corrected estimate's MSE is minimized at a 2-3 times larger bw.

I was just reminded of the time I thought I had run 1000 replications of a simulation spread across 200 processors, and as I was preparing the manuscript for submission, I realized I had actually run the same 5 replication 200 times. #youliveyoulearn

Not at #ASHG24 this year, but my excellent grad student, Dylan Steinecke, has a poster on our new sibling-based heritability estimator. Standard errors 10 times smaller than existing methods with no extra assumptions. See how on Fri, Nov 8, 2:30-4:40pm at Board 4147F!

I love the idea of the UMAP just doing the best he can with the data and assumptions he has available. Makes me want to root for him a bit more even though I know deep down he's really not up to the task. Heart of a champion!

Your first game had pictures?!?!?!