grid_by_trace: Interpolate multiple longitudinally traced patterns of...
In wfforrest/maeve: Modeling Accessories Enabling Visualization of Experiments

Description Usage Arguments Details Value Author(s) References Examples

Creates interpolated (x,y) values onto a common grid of x-values from each trace_ID, without any explicit model fit. The interpolation uses stats::approxfun() or stats::splinefun() with one of 'linear', 'trapezoid', 'step', 'spline' as its 'method', where 'linear' and 'trapezoid' are synonyms in this function. The grid onto which values are interpolated is the union of all the distinct x-values across all traces and an evenly spaced grid of 'grid_n' (see parameters) x-values spanning the x-grid range.

grid_by_trace(
  study_data_frame,
  check_data = FALSE,
  method = c("linear", "trapezoid", "step", "spline"),
  VST = TRUE,
  xmin = -Inf,
  xmax = Inf,
  xtol = (.Machine$double.eps) * 100,
  x_grid = NULL,
  grid_n = 101,
  include_obs_x = TRUE,
  subtract_starting_value = FALSE,
  na.rm = TRUE,
  singleton_to_NA = TRUE,
  group_name = maeve_options("group_name"),
  subject_ID = maeve_options("subject_ID"),
  trace_ID = maeve_options("trace_ID"),
  x_name = maeve_options("x_name"),
  endpoint_name = maeve_options("endpoint_name"),
  add_to_endpoint = maeve_options("add_to_endpoint"),
  trans_func_char = maeve_options("trans_func_char"),
  inv_func_char = maeve_options("inv_func_char"),
  test_func_x = maeve_options("test_func_x")
)

`study_data_frame`	data.frame a "well-prepared" data.frame (e.g., processed through maeve::check_study_data_frame() is recommended but not required) with the study data and appropriate columns.
`check_data`	logical whether to pass "study_data_frame" through maeve::check_study_data_frame() to check for problems.
`method`	character name method for computing AUC. Passed directly to DescTools::AUC( ..., method = ... ).
`VST`	logical whether to apply a variance stabilizing transform to endpoint_name values.
`xmin`	numeric left / lower end of the interval.
`xmax`	numeric right / uppper end of the interval.
`xtol`	machine epsilon value to decide whether an x-grid value matches one of the observed x-values.
`x_grid`	numeric vector of points at which to interpolate y-values.
`grid_n`	numeric number of points in the grid.
`include_obs_x`	logical whether to force inclusion of the union of observed 'x' times in each trace.
`subtract_starting_value`	logical whether to subtract, for each trace_ID, the first value in the time series so that each series starts at zero.
`na.rm`	logical whether to remove NA values when computing summaries.
`singleton_to_NA`	logical whether to assign trace_ID cases with a single observation NA or zero.
`group_name`	character column name for the group name factor
`subject_ID`	character column name for the subject name factor
`trace_ID`	character column name for the subject name factor
`x_name`	character column name for the x-axis / time field
`endpoint_name`	character column name for the x-axis / time field
`add_to_endpoint`	numeric offset value added to the endpoint before transformation
`trans_func_char`	character function name for transformation of (endpoint_name + add_to_endpoint)
`inv_func_char`	character inverse function for transformation of (endpoint_name + add_to_endpoint)
`test_func_x`	numeric values with which to test that trans_func and inv_func are inverse functions.

The 'trace_ID' factor can have two or more levels nested within a subject ID. Two trace levels can denote, e.g., longitudinal scans on two different tissues within the same animal, or two sequential longitudinal scans on the same tissue but with different, sequentially applied treatments.

If VST == TRUE, each ID-level vector of endpoint values (specified in 'endpoint_name') will be transformed for variance stabilization by adding a constant then transforming (e.g., "y = log( 1 + [[endpoint_name]] )"). The ID-level vector can be optionally centered to its first (in time-order) value, i.e, each y-value can have the first y-value subtracted from it.

An R data.frame with one row per ID.

Bill Forrest <forrest@gene.com>

Bill Forrest forrest@gene.com

www.r-project.org