A lot of general data from Damon Rinard's research and I devised my own very rough tests to simulate out-of-the-saddle riding at 20-30 degree bike-lean in a sprint.

1. Handlebars are easy. Just straddle the front-wheel with your legs (looking back at the bike). Grab the bar-ends and alternately push down on one end while simultaneously pull on the other. This isolates the upper-body forces from the legs. also keeps the bike-stationary for a frame of references. I can only apply about half the load of a sprint with my upper-body alone, but the bar-ends will flex close to 1" as it is.

2. wheels are very, very weak laterally. The more the bike is leaned over, the more the wheels bend at the contact patch. I clamped a wheel in Park truing stand and clamped stand in vice. Hung 100kg of sandbags off rim to simulate peak dynamic force on rim. Hang the sandbags near where the wheel goes between the arms of the stand. Measure gap between rim and arm for deflection. Used Flex=cos(LeanAngle)*MaxFlex to roughly estimate wheel-deflection where perfectly vertical bike experiences zero lateral wheel flex.

3. BB is kinda hard to isolate from the fork. Stand on left side of bike, hold front-brake, lean bike over about 20-degrees, stand on pedal with all your weight while holding same the bike-lean angle, notice BB's lateral deflection.

EDIT: Oops, wrong equation in #2, fixed.