Bug Description
When running ancombc in q2-composition, if the --p-reference-levels parameter includes only a single column::value pair, or is left blank, the tabulate visualizer produces undesirable behavior.
For the single column::value pair, the paragraph tag that contains the following text: 'Groups use to define the intercept: ...etc' produces a column separated split string for the column::value pair. Screenshot example:
For the case where --p-reference-levels is left blank, the default dummy coding column::value pair is not included in the 'Groups used to define the intercept...' tag, it is just left blank. Screenshot example:
Steps to reproduce the behavior
The example data can be used for qiime composition ancombc (single and multi formula group data can be used, the table and metadata files are the same).
To produce the examples above, either of these configurations for ancombc can be run (for single and missing reference levels, respectively):
Expected behavior
The tabulate visualizer should produce the following format for the chosen reference levels:
Single reference level:
Columns used to calculate the intercept: column1::value1 column2::value2
Multiple reference levels:
Columns used to calculate the intercept: column1::value1 column2::value2
No reference level:
Columns used to calculate the intercept: formulacolumn1::alphabetizedvalue1 formulacolumn2::alphabetizedvalue2
With formulacolumn1 and formulacolumn2 referring to the chosen column(s) from the formula parameter, and alphabetizedvalue1 and alphabetizedvalue2 referring to the default dummy coding intercept within each column (which corresponds to the highest value in alphabetical order for any categorical column).
Bug Description When running ancombc in q2-composition, if the
--p-reference-levels
parameter includes only a singlecolumn::value
pair, or is left blank, the tabulate visualizer produces undesirable behavior.For the single
column::value
pair, the paragraph tag that contains the following text: 'Groups use to define the intercept: ...etc' produces a column separated split string for thecolumn::value
pair. Screenshot example:For the case where
--p-reference-levels
is left blank, the default dummy codingcolumn::value
pair is not included in the 'Groups used to define the intercept...' tag, it is just left blank. Screenshot example:Steps to reproduce the behavior The example data can be used for
qiime composition ancombc
(single and multi formula group data can be used, the table and metadata files are the same).To produce the examples above, either of these configurations for
ancombc
can be run (for single and missing reference levels, respectively):Expected behavior The
tabulate
visualizer should produce the following format for the chosen reference levels:Single reference level: Columns used to calculate the intercept: column1::value1 column2::value2
Multiple reference levels: Columns used to calculate the intercept: column1::value1 column2::value2
No reference level: Columns used to calculate the intercept: formulacolumn1::alphabetizedvalue1 formulacolumn2::alphabetizedvalue2
With
formulacolumn1
andformulacolumn2
referring to the chosen column(s) from the formula parameter, andalphabetizedvalue1
andalphabetizedvalue2
referring to the default dummy coding intercept within each column (which corresponds to the highest value in alphabetical order for any categorical column).Computation Environment