Skip to contents

Compute contrasts with group mean imputation (DEPRECATED)

Source: R/ContrastsSimpleImpute.R
ContrastsMissing.Rd

Compute contrasts with group mean imputation (DEPRECATED)

Compute contrasts with group mean imputation (DEPRECATED)

Value

An R6 class generator.

Details

`r lifecycle::badge("deprecated")` (placeholder — prolfqua does not yet depend on the lifecycle package; consider this a soft deprecation notice. See Deprecated below.)

If there are no observations in one of the groups for some of the proteins, the group mean cannot be estimated. Therefore, assuming that the observation is missing because the protein abundance is below the detection limit, we substitute the unobserved group with the median of protein abundances observed only in one sample of the group. The variance of a protein is estimated using the pooled variance of all observations of all groups.

Deprecated

`ContrastsMissing` is a pre-fitting group-mean substitution: it does not fit a per-protein model, just stamps the group-mean delta with a pooled-variance t-test. It is superseded by build_model_impute which refits failed/singular proteins with LOD-imputed values and borrowed variance per protein. New code should prefer the lm_impute facade (ContrastsLMImputeFacade) or any of its limma_impute / limma_voom_impute cousins. Constructing ContrastsMissing emits a .Deprecated warning at initialize time.

Still reachable via model = "lm_missing" for users who want to explicitly run the group-mean fallback; the construction emits a .Deprecated warning each time. Will be removed once a merge-style replacement (e.g. an lm_impute_missing facade composing build_model with build_model_impute) lands.

See also

build_model_impute, ContrastsLMImputeFacade

Other modelling: AnovaExtractor, Contrasts, ContrastsDEqMSFacade, ContrastsDEqMSVoomFacade, ContrastsFirth, ContrastsFirthFacade, ContrastsFirthNestedFacade, ContrastsLMFacade, ContrastsLMImputeFacade, ContrastsLMMissingFacade, ContrastsLimma, ContrastsLimmaFacade, ContrastsLimmaImputeFacade, ContrastsLimmaVoomFacade, ContrastsLimmaVoomImputeFacade, ContrastsLimpaFacade, ContrastsLimpaNestedFacade, ContrastsLmerNestedFacade, ContrastsModerated, ContrastsModeratedDEqMS, ContrastsPlotter, ContrastsRLMFacade, ContrastsROPECA, ContrastsROPECANestedFacade, ContrastsTable, INTERNAL_FUNCTIONS_BY_FAMILY, LR_test(), Model, ModelFirth, ModelLimma, StrategyLM, StrategyLimma, StrategyLimpa, StrategyLmer, StrategyLogistf, StrategyRLM, build_contrast_analysis(), build_model(), build_model_glm_peptide(), build_model_glm_protein(), build_model_impute(), build_model_limma(), build_model_limma_impute(), build_model_limma_voom(), build_model_limma_voom_impute(), build_model_limpa(), build_model_logistf(), compute_borrowed_variance(), compute_borrowed_variance_limma(), compute_contrast(), compute_lmer_contrast(), contrasts_fisher_exact(), get_anova_df(), get_complete_model_fit(), get_p_values_pbeta(), group_label(), impute_refit_singular(), is_singular_lm(), linfct_all_possible_contrasts(), linfct_factors_contrasts(), linfct_from_model(), linfct_matrix_contrasts(), list_facades(), lookup_facade(), merge_contrasts_results(), model_analyse(), model_summary(), moderated_p_deqms(), moderated_p_deqms_long(), moderated_p_limma(), moderated_p_limma_long(), new_lm_imputed(), pivot_model_contrasts_to_wide(), plot_lmer_peptide_predictions(), register_facade(), sim_build_models_lm(), sim_build_models_lmer(), sim_build_models_logistf(), sim_make_model_lm(), sim_make_model_lmer(), strategy_limma(), strategy_limpa(), strategy_logistf(), summary_ROPECA_median_p.scaled(), unregister_facade()

Super class

prolfqua::ContrastsInterface -> ContrastsMissing

Public fields

subject_id

subject_id e.g. protein_ID column

contrasts

array with contrasts (see example)

model_name

model name

contrast_result

data frame with results of contrast computation

lfqdata

data frame

confint

confidence interval

p.adjust

function to adjust p-values

global

Take global or local values for imputation

present

default 1, presence in interaction to infer limit of detection.

minsd

default 1, if standard deviation can not be estimated, what is the prior minimum sd, default = 1s

Methods

Public methods

Inherited methods

Method new()

initialize

Usage

ContrastsMissing$new(
  lfqdata,
  contrasts,
  confint = 0.95,
  p.adjust = prolfqua::adjust_p_values,
  model_name = "groupAverage"
)

Arguments

lfqdata

LFQData

contrasts

array of contrasts (see example)

confint

confidence interval

p.adjust

method for p-value adjustment - default Benjamini Hochberg

model_name

default "groupAverage"

Method get_contrast_sides()

get contrasts sides

Usage

ContrastsMissing$get_contrast_sides()

Method get_contrasts()

table with results of contrast computation

Usage

ContrastsMissing$get_contrasts(all = FALSE)

Arguments

all

FALSE, do not show all columns (default)

Method get_Plotter()

get ContrastsPlotter

Usage

ContrastsMissing$get_Plotter()

Returns

Contrast_Plotter

Method to_wide()

convert contrast results to wide format

Usage

ContrastsMissing$to_wide(columns = c("p.value", "FDR", "statistic"))

Arguments

columns

value column default p.value

Returns

data.frame

Method clone()

The objects of this class are cloneable with this method.

Usage

ContrastsMissing$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

Nprot <- 120
istar <- prolfqua::sim_lfq_data_protein_config(Nprot = Nprot,weight_missing = .4)
#> creating sampleName from file_name column
#> completing cases
#> completing cases done
#> setup done
istar$data$abundance |> is.na() |> sum()
#> [1] 283
protIntensity <- istar$data
config <- istar$config


lProt <- LFQData$new(protIntensity, config)
lProt$rename_response("transformedIntensity")

Contr <- c("dil.b_vs_a" = "group_A - group_Ctrl")
csi <- ContrastsMissing$new(lProt, contrasts = Contr)
#> Warning: ContrastsMissing is deprecated: it substitutes group means rather than fitting a model. Prefer build_model_impute (LOD-imputed per-protein refit with borrowed variance) via the lm_impute / limma_impute facades. See ?ContrastsMissing for details.
csi$get_contrast_sides()
#> # A tibble: 1 × 3
#>   contrast   group_1 group_2   
#>   <chr>      <chr>   <chr>     
#> 1 dil.b_vs_a group_A group_Ctrl

res <- csi$get_contrasts()
#> dil.b_vs_a=group_A - group_Ctrl
#> dil.b_vs_a=group_A - group_Ctrl
#> dil.b_vs_a=group_A - group_Ctrl

stopifnot(nrow(res) ==  (protIntensity$protein_Id |> unique() |> length()))
res$contrast |> table()
#> 
#> dil.b_vs_a 
#>        120 
stopifnot((res$p.value |> is.na() |> sum()) == 0)
plot(res$diff, -log10(res$p.value), pch = ".")

csi$column_description()
#>           column_name
#> modelName   modelName
#> contrast     contrast
#> avgAbd         avgAbd
#> diff             diff
#> FDR               FDR
#> statistic   statistic
#> std.error   std.error
#> df                 df
#> p.value       p.value
#> conf.low     conf.low
#> conf.high   conf.high
#> sigma           sigma
#>                                                                                            description
#> modelName                                                                                type of model
#> contrast                                                      name of difference e.g. group1_vs_group2
#> avgAbd                                                  mean abundance value of protein in all samples
#> diff                                                                       difference among conditions
#> FDR                                                                               false discovery rate
#> statistic                                                                                 t-statistics
#> std.error                                                                               standard error
#> df                                                                                  degrees of freedom
#> p.value                                                                                        p-value
#> conf.low                                                         lower value of 95 confidence interval
#> conf.high                                                         high value of 95 confidence interval
#> sigma     residual standard deviation of linear model (needed for empirical Bayes variance shrinkage).
x<- csi$get_Plotter()
p <- x$volcano()
pdf(file = NULL)
print(p)
#> $p.value
#> 
#> $FDR
#> 
dev.off()
#> agg_record_1bde3841b69 
#>                      2 

dd <- prolfqua::sim_lfq_data_2factor_config(Nprot = 100,weight_missing = 0.1)
#> creating sampleName from file_name column
#> completing cases
#> completing cases done
#> setup done

Contrasts <- c("c1" = "TreatmentA - TreatmentB",
               "C2" = "BackgroundX- BackgroundZ",
               "c3" = "`TreatmentA:BackgroundX` - `TreatmentA:BackgroundZ`",
               "c4" = "`TreatmentB:BackgroundX` - `TreatmentB:BackgroundZ`"
               )
lProt <- LFQData$new(dd$data, dd$config)
lProt$rename_response("transformedIntensity")

csi <- ContrastsMissing$new(lProt, contrasts = Contrasts)
#> Warning: ContrastsMissing is deprecated: it substitutes group means rather than fitting a model. Prefer build_model_impute (LOD-imputed per-protein refit with borrowed variance) via the lm_impute / limma_impute facades. See ?ContrastsMissing for details.
res <- csi$get_contrasts()
#> c1=TreatmentA - TreatmentB
#> C2=BackgroundX- BackgroundZ
#> c3=`TreatmentA:BackgroundX` - `TreatmentA:BackgroundZ`
#> c4=`TreatmentB:BackgroundX` - `TreatmentB:BackgroundZ`
#> c1=TreatmentA - TreatmentB
#> C2=BackgroundX- BackgroundZ
#> c3=`TreatmentA:BackgroundX` - `TreatmentA:BackgroundZ`
#> c4=`TreatmentB:BackgroundX` - `TreatmentB:BackgroundZ`
#> c1=TreatmentA - TreatmentB
#> C2=BackgroundX- BackgroundZ
#> c3=`TreatmentA:BackgroundX` - `TreatmentA:BackgroundZ`
#> c4=`TreatmentB:BackgroundX` - `TreatmentB:BackgroundZ`
pl <- csi$get_Plotter()
pdf(file = NULL)
pl$volcano()
#> $p.value
#> 
#> $FDR
#> 
dev.off()
#> agg_record_1bde3841b69 
#>                      2