Skip to contents

expression() is an analysis function that computes the protein summary statistics for a given tidyproteomics data object.

Usage

expression(
  data = NULL,
  ...,
  .pairs = NULL,
  .method = stats::t.test,
  .p.adjust = "BH"
)

Arguments

data

tidyproteomics data object

...

two sample comparison e.g. experimental/control

.method

a two-distribution test function returning a p_value for the null hypothesis. Example functions include t.test, wilcox.test, stats::ks.test, additionally, the string "limma" can be used to select from the limma package to compute an empirical Bayesian estimation which performs better with non-linear distributions and uneven replicate balance between samples.

.p.adjust

a stats::p.adjust string for multiple test correction, default is 'BH' (Benjamini & Hochberg, 1995)

Value

a tibble

Examples

library(dplyr, warn.conflicts = FALSE)
library(tidyproteomics)

# simple t.test expression analysis
hela_proteins %>%
   expression(knockdown/control) %>%
   export_analysis(knockdown/control, .analysis = "expression")
#>  .. expression::t_test testing knockdown / control
#>  .. expression::t_test testing knockdown / control [3.4s]
#> 
#> # A tibble: 5,187 × 27
#>    protein imputed     n imputed_control imputed_knockdown average_expression
#>    <chr>     <dbl> <int>           <dbl>             <dbl>              <dbl>
#>  1 Q5SVJ8    0.333     6           0.667                 0            755886.
#>  2 Q15011    0         6           0                     0           4919332.
#>  3 F6SA91    0.167     6           0.333                 0            879218.
#>  4 Q9NXV2    0         6           0                     0           3103100.
#>  5 Q13751    0         6           0                     0           1072517.
#>  6 O00308    0         6           0                     0            733272.
#>  7 Q9BY50    0         6           0                     0           2049680.
#>  8 Q8TBM8    0.333     6           0.667                 0           3374207.
#>  9 Q5JUW8    0         6           0                     0           1745373.
#> 10 Q5T9L3    0.167     6           0.333                 0            961926.
#> # ℹ 5,177 more rows
#> # ℹ 21 more variables: proportional_expression <dbl>, foldchange <dbl>,
#> #   log2_foldchange <dbl>, p_value <dbl>, adj_p_value <dbl>,
#> #   normalization <chr>, abundance_control_1 <dbl>, abundance_control_2 <dbl>,
#> #   abundance_control_3 <dbl>, abundance_knockdown_1 <dbl>,
#> #   abundance_knockdown_2 <dbl>, abundance_knockdown_3 <dbl>,
#> #   description <chr>, biological_process <chr>, cellular_component <chr>, …

# a wilcox.test expression analysis
hela_proteins %>%
   expression(knockdown/control, .method = stats::wilcox.test) %>%
   export_analysis(knockdown/control, .analysis = "expression")
#>  .. expression::wilcox_test testing knockdown / control
#>  .. expression::wilcox_test testing knockdown / control [3.1s]
#> 
#> # A tibble: 5,187 × 27
#>    protein imputed     n imputed_control imputed_knockdown average_expression
#>    <chr>     <dbl> <int>           <dbl>             <dbl>              <dbl>
#>  1 Q5SVJ8    0.333     6           0.667                 0            755886.
#>  2 Q15011    0         6           0                     0           4919332.
#>  3 F6SA91    0.167     6           0.333                 0            879218.
#>  4 Q9NXV2    0         6           0                     0           3103100.
#>  5 Q13751    0         6           0                     0           1072517.
#>  6 O00308    0         6           0                     0            733272.
#>  7 Q9BY50    0         6           0                     0           2049680.
#>  8 Q8TBM8    0.333     6           0.667                 0           3374207.
#>  9 Q5JUW8    0         6           0                     0           1745373.
#> 10 Q5T9L3    0.167     6           0.333                 0            961926.
#> # ℹ 5,177 more rows
#> # ℹ 21 more variables: proportional_expression <dbl>, foldchange <dbl>,
#> #   log2_foldchange <dbl>, p_value <dbl>, adj_p_value <dbl>,
#> #   normalization <chr>, abundance_control_1 <dbl>, abundance_control_2 <dbl>,
#> #   abundance_control_3 <dbl>, abundance_knockdown_1 <dbl>,
#> #   abundance_knockdown_2 <dbl>, abundance_knockdown_3 <dbl>,
#> #   description <chr>, biological_process <chr>, cellular_component <chr>, …

# a one-tailed wilcox.test expression analysis
wilcoxon_less <- function(x, y) {
   stats::wilcox.test(x, y, alternative = "less")
}
hela_proteins <- hela_proteins %>%
   expression(knockdown/control, .method = stats::wilcox.test)
#>  .. expression::wilcox_test testing knockdown / control
#>  .. expression::wilcox_test testing knockdown / control [3.2s]
#> 

hela_proteins %>% export_analysis(knockdown/control, .analysis = "expression")
#> # A tibble: 5,187 × 27
#>    protein imputed     n imputed_control imputed_knockdown average_expression
#>    <chr>     <dbl> <int>           <dbl>             <dbl>              <dbl>
#>  1 Q5SVJ8    0.333     6           0.667                 0            755886.
#>  2 Q15011    0         6           0                     0           4919332.
#>  3 F6SA91    0.167     6           0.333                 0            879218.
#>  4 Q9NXV2    0         6           0                     0           3103100.
#>  5 Q13751    0         6           0                     0           1072517.
#>  6 O00308    0         6           0                     0            733272.
#>  7 Q9BY50    0         6           0                     0           2049680.
#>  8 Q8TBM8    0.333     6           0.667                 0           3374207.
#>  9 Q5JUW8    0         6           0                     0           1745373.
#> 10 Q5T9L3    0.167     6           0.333                 0            961926.
#> # ℹ 5,177 more rows
#> # ℹ 21 more variables: proportional_expression <dbl>, foldchange <dbl>,
#> #   log2_foldchange <dbl>, p_value <dbl>, adj_p_value <dbl>,
#> #   normalization <chr>, abundance_control_1 <dbl>, abundance_control_2 <dbl>,
#> #   abundance_control_3 <dbl>, abundance_knockdown_1 <dbl>,
#> #   abundance_knockdown_2 <dbl>, abundance_knockdown_3 <dbl>,
#> #   description <chr>, biological_process <chr>, cellular_component <chr>, …

# Note: the userdefined function is preserved in the operations tracking
hela_proteins %>% operations()
#>  Data Transformations
#>   • Data files (p97KD_HCT116_proteins.xlsx) were imported as proteins from
#>   ProteomeDiscoverer
#>   • Analysis: expression difference wilcox_test knockdown/control, p.adjust =
#>   BH

# limma expression analysis
hela_proteins %>%
   expression(knockdown/control, .method = "limma") %>%
   export_analysis(knockdown/control, .analysis = "expression")
#>  .. expression::limma testing knockdown / control
#> ! expression::limma removed 159 proteins with completely missing values
#>  .. expression::limma testing knockdown / control

#>  .. expression::limma testing knockdown / control [564ms]
#> 
#> # A tibble: 5,187 × 29
#>    protein imputed     n imputed_control imputed_knockdown log2_foldchange
#>    <chr>     <dbl> <int>           <dbl>             <dbl>           <dbl>
#>  1 Q15011    0         6           0                     0            5.48
#>  2 F6SA91    0.167     6           0.333                 0            5.31
#>  3 Q5SVJ8    0.333     6           0.667                 0            3.99
#>  4 Q13751    0         6           0                     0            3.66
#>  5 Q9NXV2    0         6           0                     0            3.63
#>  6 O00308    0         6           0                     0            3.34
#>  7 Q9BY50    0         6           0                     0            3.09
#>  8 Q5JUW8    0         6           0                     0            3.05
#>  9 Q5T9L3    0.167     6           0.333                 0            2.99
#> 10 O95833    0.167     6           0.333                 0            2.88
#> # ℹ 5,177 more rows
#> # ℹ 23 more variables: foldchange <dbl>, average_expression <dbl>,
#> #   proportional_expression <dbl>, p_value <dbl>, adj_p_value <dbl>,
#> #   limma_t_statistic <dbl>, limma_B_statistic <dbl>, normalization <chr>,
#> #   abundance_control_1 <dbl>, abundance_control_2 <dbl>,
#> #   abundance_control_3 <dbl>, abundance_knockdown_1 <dbl>,
#> #   abundance_knockdown_2 <dbl>, abundance_knockdown_3 <dbl>, …

# using the .pairs argument when multiple comparisons are needed
comps <- list(c("control","knockdown"),
            c("knockdown","control"))

hela_proteins %>%
   expression(.pairs = comps)
#> Expression Analysis - using the supplied 2 sample pairs ...
#>  .. expression::t_test testing control / knockdown
#>  .. expression::t_test testing control / knockdown [3.3s]
#> 
#>  .. expression::t_test testing knockdown / control
#>  .. expression::t_test testing knockdown / control [3.2s]
#> 
#> 
#> ── Quantitative Proteomics Data Object ──
#> 
#> Origin          ProteomeDiscoverer 
#>                 proteins (12.13 MB) 
#> Composition     6 files 
#>                 2 samples (control, knockdown) 
#> Quantitation    7055 proteins 
#>                 4 log10 dynamic range 
#>                 28.8% missing values 
#>  *imputed        
#> Accounting      (4) num_peptides num_psms num_unique_peptides imputed 
#> Annotations     (9) description biological_process cellular_component molecular_function
#>                 gene_id_entrez gene_name wiki_pathway reactome_pathway
#>                 gene_id_ensemble 
#> Analyses        (2) 
#>                 knockdown/control -> expression  
#>                 control/knockdown -> expression  
#>