Partition Trait into Signal, Rest and Noise

Partition a trait value into three parts based on rest and signal factors. The signal is the main interest for comparison across traits, while the rest concerns aspects of the experiment that are controlled. Noise is unexplained variation. The three parts will be orthogonal to each other. The `signal` factors, entered as terms for `formula`, are conditioned by the `rest` factors.

Usage

partition(
  object,
  trait = "trait",
  value = "value",
  signal = ifelse(is_condition, "strain * sex * condition", "strain * sex"),
  rest = ifelse(is_condition, "strain * sex + sex * condition", "sex")
)

Arguments

object

data frame in long format with trait data

trait

name of column with trait names

value

name column with trait values

signal

signal factor combination as string for `formula`

rest

rest factor combination as string for `formula`

Value

data frame with added columns `signal`, `cellmean`

Examples

partition(sampleData)
#> # A tibble: 128 × 7
#>    dataset trait strain sex   condition  signal cellmean
#>    <chr>   <fct> <chr>  <chr> <chr>       <dbl>    <dbl>
#>  1 sample  C     B6     F     Y          0.311    0.599 
#>  2 sample  C     B6     F     X         -0.414   -0.0951
#>  3 sample  C     B6     M     X         -0.429   -1.62  
#>  4 sample  C     B6     M     Y          0.286   -1.00  
#>  5 sample  C     129    M     X         -0.0601  -0.648 
#>  6 sample  C     129    F     X          0.0751  -0.307 
#>  7 sample  C     129    F     Y         -0.0751  -0.487 
#>  8 sample  C     129    M     Y          0.0601  -0.626 
#>  9 sample  C     AJ     F     X          0.445    1.67  
#> 10 sample  C     AJ     F     Y         -0.445    0.752 
#> # ℹ 118 more rows