A two-fer on the decline effect today (or perhaps this is less decline effect than just the annals of irreproducible results): when studying a microdeletion in the 16p12.1 region of the human genome, researchers learned that the published human genome was incorrect.  That's got to be frustrating.