explaining "study_condition" and "disease" fields

[schifferl]

From @edoardopasolli on December 7, 2017 19:50

An external user may have some difficulties in understanding how we defined the "study_condition" and "disease" fields. Should we clarify this aspect better somewhere?

Copied from original issue: waldronlab/curatedMetagenomicData#118

[schifferl]

From @lwaldron on December 8, 2017 21:30

I think the canonical location for variable documentation is in the "Description" column of https://github.com/waldronlab/curatedMetagenomicDataCuration/blob/master/inst/extdata/template.csv. If you would add this @edoardopasolli (and check the adequacy of the rest of the descriptions), then we'll convert this issue to one for @schifferl to parse this file to create automatic codebooks in the vignette and in the combined_metadata help page.

[lwaldron]

• study_condition is always one word, and is the label for this participant for the purposes of the analysis this study was designed around. In case-control studies, the controls are labeled "control". In studies without a contrast, all participants are labeled "control". The main comparison group is labeled according to what that condition or exposure is.

• disease can be a semicolon-delimited array of diseases associated with the participant, even if they are not under study. "none" if no disease is known for the participant.

• annotate the design of each an array of study designs, consistently with bugsigdb.org:

  • case-control
  • cross-sectional observational, not case-control
  • time series / longitudinal observational
  • laboratory experiment
  • randomized controlled trial
  • prospective cohort

[lwaldron]

I think the main action item here is for @QuanWan89 to include a study design field in the study table in the new metadata database, with the following values. That can be started in Quan's project management (although I don't have access to that):

• case-control
• cross-sectional observational, not case-control
• time series / longitudinal observational
• laboratory experiment
• randomized controlled trial
• prospective cohort