ref: peds-sim

[cherman2]

we need to refactor peds simulation to have an even sampling depth

Peds-Simulation Process:

1. create mismatched pairs
2. sample mismatched pairs so there are 999 mismatched pairs (simulated mismatched pairs)
3. create a feature table that evenly samples the donor and recipient from the pair (but does this independently for all pairs in the simulated mismatch pairs data)
4. create new mismatched pairs because we are people going to have to rename sample for the sampling iteration they are a part of
5. calculate simulated peds as normal
6. compare all real values to the new simulated peds distribution

Peds sampling depth process:
For this one we are going to default to none but will allow sampling if a sampling depth is provided and will allow rarefaction if iters is provided.

1. sample feature table 999 times
2. run peds 999 times
3. Average peds values

Alright, (@gregcaporaso if you are interested) This are my rough note from what Evan and I decided re: peds-sim and peds sampling depth.
@ebolyen Does this seem right to you?