Random forest eigen values

[wagenrace]

The following pull request adds, fixes, and/or modifies:

Explored eigen values with random forests.
@gwaygenomics and @koalive, I really can use your help here. The results get worst by adding eigen values to the mix even if I only pick the ones with a significant different in the targets.
But I cannot really explain why.
So far I have 2 hypothesizes:

• There is a bug some where
• The extra features are still to noisy making it so the trees overfit

I performed the following prior to filing this pull request:

• Tested that my change does not break the analysis pipeline
• Ran a linter through my code
• Update environment dependencies if my code introduces a new package