62 bidirectional summary stats

[cthunes]

Added the following eight summary stats

• min_mean - the minimum mean response (interpretable as the max mean response in the negative direction)
• min_mean_conc - the concentration at which the min_mean was observed
• min_med - the minimum median response (interpretable as the max median response in the negative direction)
• min_med_conc - the concentration at which the min_med was observed
• max_med_diff - the median response furthest from baseline, or zero, regardless of direction (the maximum median difference). Can be negative or positive.
• max_med_diff_conc - the concentration at which the max_med_diff was observed
• nmed_gtbl_pos - the number of median responses greater than baseline (3 * bmad) in the positive direction. Same implementation as nmed_gtbl. nmed_gtbl values have been copied to nmed_gtbl_pos.
  • nmed_gtbl will need to be dropped at a later date while refactoring any code that uses it, may need another ticket.
• nmed_gtbl_neg - the number of median responses greater than baseline (-3 * bmad) in the negative direction

Table schema for mc4 has been updated to reflect these new summary stats. Uploaded file to Clowder and tracked in schema_changes table. Closes #62.

[madison-feshuk]

Saw following error writing Level 4:

Writing level 4 data for 4 ids...
Completed delete cascade for 4 ids (4.94 secs)
Error in h(simpleError(msg, call)) :
error in evaluating the argument 'value' in selecting a method for function 'dbWriteTable': Some items of .SDcols are not column names: [nmed_gtbl]

[madison-feshuk]

Looks like nmed_gtbl might be dropped in code, but column is kept in the database

[cthunes]

Column has been dropped in invitrodb

[cthunes]

Writing and loading data now accounting for new columns. I didn't change the columns for code using old schema, is this assumption correct?

[sedavid01]

@cthunes, was this changed from 'bidirectional' passed argument to TRUE since this the max_med_diff and max_med_diff_conc are only apply to the bidirectional = TRUE cases for right now?

[sedavid01]

Nevermind.... This is because we are only fitting in the bi-directional direction now within tcpl.

[cthunes]

Including updated mc6 methods that formerly used nmed_gtbl in this PR. Closes #127

[madison-feshuk]

Looks good with padilla processing! This is a huge update. I think we should merge this branch to dev, where we can continue testing