Reorganize motif half life model code?

[ewallace]

Currently:

• train_half_life_linear_model.R organizes the data and then does the training/ motif search, that takes several minutes to run.
• and train_half_life_linear_model.R also outputs some plot panels.
• all those outputs go through a .Rdata binary file that isn't committed to the repo.
• hlife_model_extract_motif_coef.R does a mixture of new linear modeling for selected coefficients, making a new coefficient plot, and assembling a multipanel plot.

This makes it difficult and time-consuming to edit the figure formatting by itself. And hard to follow.

Can we make this easier at all? Separate the plotting from the calculation, or pass the data round more nicely?

[ewallace]

Maybe if train_half_life_linear_model.R could output a flat table combined_hlife_data_sets, containing everything needed for the figures. Also a flat table (comibned?) of coefficients. Then we could load and use those for building figures in the other script.