DPA will tell me about how much time is spent waiting for memory allocations (although only aggregated with other memory and CPU waits), but that doesn't tell me which statements executing during a time period used the most buffer cache, which is generally what causes a Page Life Expectancy drop.
If I drill into the minutes in between the last good PLE reading and the first very low one, then select the Waits tab and pick Memory/CPU, and then look at the top SQL, none of them are ever the cause of the PLE drop.
