adding sequential all to top method

[cmatKhan]

yeastdnnexplorer/yeastdnnexplorer/__main__.py

Line 223 in a72285f

I think if we do the "all" result from lasso res outside of the dict, and then take the sig predictors from that, we can subset the X with those sig predictors and then run the lasso bootstrap function using that subset of X.

That would then get set into lasso_res as "top" and then everything else would be the same.

I think that argues for a function to generate the lasso_res with an if/else to set the method (something like "all_top" or "all_sequential_top")

-- the goal here is to do as little as possible. I think lasso_res['all'] is the one that must exist after the "connection point" where "final_coefficient_set" (not the right term, but something like that) is created. So, as long as lasso_res['all'] exists, we shouldn't have to change anything below that point (i think).