Comparing Two Groups with prolfqua

Purpose

This vignette demonstrates how two conditions, e.g., treatment vs. control, can be compared and the differences statistically tested. We will again use the Ionstar dataset as an example of an LFQ experiment. This dataset was preprocessed with the MaxQuant software. After first examining the data using QC plots and then normalizing the data, we compare groups of replicates with the different dilutions. The output of the comparison is the difference in the mean intensities for quantified proteins (log2 fold-change) in each group, along with statistical parameters such as degrees of freedom, standard errors, p-value and the FDR.

Loading protein abundances from MaxQuant proteinGroups.txt

library(prolfqua)

Specify the path to the MaxQuant proteinGroups.txt file. The function tidyMQ_ProteinGroups will read the proteinGroups.txt file and convert it into a tidy table

xx <- prolfqua::sim_lfq_data_protein_config(Nprot = 100)
xx

## $data
## # A tibble: 1,200 × 8
##    sample sampleName group_ isotopeLabel protein_Id abundance qValue nr_peptides
##    <chr>  <chr>      <chr>  <chr>        <chr>          <dbl>  <dbl>       <dbl>
##  1 A_V1   A_V1       A      light        0EfVhX~39…      20.4      0           2
##  2 A_V1   A_V1       A      light        0m5WN4~60…      16.4      0           7
##  3 A_V1   A_V1       A      light        0YSKpy~28…      18.2      0           2
##  4 A_V1   A_V1       A      light        3QLHfm~89…      23.0      0           1
##  5 A_V1   A_V1       A      light        3QYop0~75…      23.1      0           7
##  6 A_V1   A_V1       A      light        76k03k~70…      25.5      0           4
##  7 A_V1   A_V1       A      light        7cbcrd~73…      27.1      0           2
##  8 A_V1   A_V1       A      light        7QuTub~18…      18.4      0           4
##  9 A_V1   A_V1       A      light        7soopj~53…      19.5      0           2
## 10 A_V1   A_V1       A      light        7zeekV~71…      20.6      0           8
## # ℹ 1,190 more rows
## 
## $config
## <AnalysisConfiguration>
##   Public:
##     clone: function (deep = FALSE) 
##     initialize: function (analysisTableAnnotation, analysisParameter = AnalysisParameters$new()) 
##     parameter: AnalysisParameters, R6
##     sep: ~
##     table: AnalysisTableAnnotation, R6

Create the LFQData class instance and remove zeros from data (MaxQuant encodes missing values with zero).

lfqdata <- prolfqua::LFQData$new(xx$data, xx$config)
lfqdata$remove_small_intensities()
lfqdata$factors()

## # A tibble: 12 × 3
##    sample  sampleName group_
##    <chr>   <chr>      <chr> 
##  1 A_V1    A_V1       A     
##  2 A_V2    A_V2       A     
##  3 A_V3    A_V3       A     
##  4 A_V4    A_V4       A     
##  5 B_V1    B_V1       B     
##  6 B_V2    B_V2       B     
##  7 B_V3    B_V3       B     
##  8 B_V4    B_V4       B     
##  9 Ctrl_V1 Ctrl_V1    Ctrl  
## 10 Ctrl_V2 Ctrl_V2    Ctrl  
## 11 Ctrl_V3 Ctrl_V3    Ctrl  
## 12 Ctrl_V4 Ctrl_V4    Ctrl

You can always convert the data into wide format.

lfqdata$to_wide()$data[1:3,1:7]

## # A tibble: 3 × 7
##   protein_Id  isotopeLabel  A_V1  A_V2  A_V3  A_V4  B_V1
##   <chr>       <chr>        <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 0EfVhX~3967 light         20.4  23.8  22.3  21.5  19.1
## 2 0m5WN4~6025 light         16.4  NA    16.1  18.2  24.3
## 3 0YSKpy~2865 light         18.2  17.7  16.7  17.7  18.2

Visualization of not normalized data

After this first setting up of the analysis we show now how to normalize the proteins and the effect of normalization. Furthermore we use some functions to visualize the missing values in our data.

lfqplotter <- lfqdata$get_Plotter()
density_nn <- lfqplotter$intensity_distribution_density()

Visualization of missing data

lfqplotter$NA_heatmap()

Heatmap where missing proteins (zero in case of MaxQuant reported intensities), black - missing protein intensities, white - present

lfqdata$get_Summariser()$plot_missingness_per_group()

# of proteins with 0,1,…N missing values

lfqplotter$missigness_histogram()

Intensity distribution of proteins depending on # of missing values

Computing standard deviations, mean and CV.

Other important statistics such as coefficient of variation, means and standard deviations can be easily calculated using the get_Stats function and visualized with a violin plot.

stats <- lfqdata$get_Stats()

## [1] "group_"
## NULL

stats$violin()

Violin plots of CVs in the different groups and among all groups

prolfqua::table_facade( stats$stats_quantiles()$wide, paste0("quantile of ",stats$stat ))

quantile of CV
probs	A	All	B	Ctrl
0.10	2.014611	3.451035	1.993591	1.762209
0.25	2.711702	4.370724	2.956419	2.882970
0.50	3.655067	5.296948	3.700140	4.261204
0.75	5.338899	7.411380	5.462486	5.904953
0.90	7.541564	9.912513	7.118907	7.471854

stats$density_median()

Distribution of CV’s for top 50% and bottom 50% proteins by intensity.

Normalize protein intensities and show diagnostic plots

We normalize the data by \(\log_2\) transforming and then \(z-scaling\).

lt <- lfqdata$get_Transformer()
transformed <- lt$log2()$robscale()$lfq
transformed$config$table$is_response_transformed

## [1] TRUE

pl <- transformed$get_Plotter()
density_norm <- pl$intensity_distribution_density()

gridExtra::grid.arrange(density_nn, density_norm)

Distribution of intensities before and after normalization.

pl$pairs_smooth()

Scatterplot matrix

## NULL

p <- pl$heatmap_cor()

Heatmap, Rows - proteins, Columns - samples

Fitting a linear model

For fitting linear models to the transformed intensities for all our proteins we have to first specify the model function and define the contrasts that we want to calculate.

transformed$factors()

## # A tibble: 12 × 3
##    sample  sampleName group_
##    <chr>   <chr>      <chr> 
##  1 A_V1    A_V1       A     
##  2 A_V2    A_V2       A     
##  3 A_V3    A_V3       A     
##  4 A_V4    A_V4       A     
##  5 B_V1    B_V1       B     
##  6 B_V2    B_V2       B     
##  7 B_V3    B_V3       B     
##  8 B_V4    B_V4       B     
##  9 Ctrl_V1 Ctrl_V1    Ctrl  
## 10 Ctrl_V2 Ctrl_V2    Ctrl  
## 11 Ctrl_V3 Ctrl_V3    Ctrl  
## 12 Ctrl_V4 Ctrl_V4    Ctrl

formula_Condition <-  strategy_lm("transformedIntensity ~ group_")

# specify model definition
modelName  <- "Model"
Contrasts <- c("AvsC" = "group_A - group_Ctrl",
"BvsC" = "group_B - group_Ctrl")

Here we have to build the model for each protein.

mod <- prolfqua::build_model(
  transformed$data,
  formula_Condition,
  subject_Id = transformed$config$table$hierarchy_keys() )

In this plot we can see what factors in our model are mostly responsible for the adjusted p-values calculated from an analysis of variance.

mod$anova_histogram("FDR")

## $plot

Distribtuion of adjusted p-values (FDR)

## 
## $name
## [1] "Anova_p.values_Model.pdf"

One also look what proteins do show different abundances in any of our five dilutions by looking at the FDR values of the analysis of variane.

aovtable <- mod$get_anova()
head(aovtable)

## # A tibble: 6 × 10
##   protein_Id  isSingular nrcoef factor    Df  Sum.Sq  Mean.Sq F.value  p.value
##   <chr>       <lgl>       <int> <chr>  <int>   <dbl>    <dbl>   <dbl>    <dbl>
## 1 0EfVhX~3967 FALSE           3 group_     2 0.0626  0.0313     5.18  0.0416  
## 2 0m5WN4~6025 FALSE           3 group_     2 0.324   0.162     29.4   0.000205
## 3 0YSKpy~2865 FALSE           3 group_     2 0.00753 0.00376    0.605 0.569   
## 4 3QLHfm~8938 FALSE           3 group_     2 0.0637  0.0319    13.2   0.00213 
## 5 3QYop0~7543 FALSE           3 group_     2 0.00121 0.000603   0.178 0.840   
## 6 76k03k~7094 FALSE           3 group_     2 0.0264  0.0132     3.06  0.0968  
## # ℹ 1 more variable: FDR <dbl>

dim(aovtable)

## [1] 98 10

xx <- aovtable |> dplyr::filter(FDR < 0.2)
signif <- transformed$get_copy()
signif$data <- signif$data |> dplyr::filter(protein_Id %in% xx$protein_Id)
hmSig <- signif$get_Plotter()$heatmap()

hmSig

Heatmap for proteins with FDR < 0.2 in the analysis of variance

Compute contrasts

Next we do calculate the statistics for our defined contrasts for all the proteins. For this we can use the Contrasts function.

contr <- prolfqua::Contrasts$new(mod, Contrasts)
v1 <- contr$get_Plotter()$volcano()

Alternatively, we can moderate the variance and using the Experimental Bayes method implemented in ContrastsModerated.

contr <- prolfqua::ContrastsModerated$new(contr)
contrdf <- contr$get_contrasts()

In the next figure it can be seen WEWinputNEEDED.

plotter <- contr$get_Plotter()
v2 <- plotter$volcano()
gridExtra::grid.arrange(v1$FDR,v2$FDR, ncol = 1)

Volcano plot, Left panel - no moderation, Right panel - with moderation.

plotter$ma_plotly()

MA plot showing the dependency of mean abuncance with respect to the difference

#myProteinIDS <- c("sp|Q12246|LCB4_YEAST",  "sp|P38929|ATC2_YEAST",  "sp|Q99207|NOP14_YEAST")
myProteinIDS <- c("sp|P0AC33|FUMA_ECOLI",  "sp|P28635|METQ_ECOLI",  "sp|Q14C86|GAPD1_HUMAN")
dplyr::filter(contrdf, protein_Id %in% myProteinIDS)

## # A tibble: 0 × 13
## # ℹ 13 variables: modelName <chr>, protein_Id <chr>, contrast <chr>,
## #   diff <dbl>, std.error <dbl>, avgAbd <dbl>, statistic <dbl>, df <dbl>,
## #   p.value <dbl>, conf.low <dbl>, conf.high <dbl>, sigma <dbl>, FDR <dbl>

Contrasts with missing value imputation

Use this method if there proteins with no observations in one of the groups. With the ContrastsMissing function, we can estimate difference in mean for proteins that are not observed in one group or condition. For this we are using the average expression at percentile 0.05 of the group where the protein is not quantified.

mC <- ContrastsMissing$new(lfqdata = transformed, contrasts = Contrasts)
colnames(mC$get_contrasts())

## [1] "group_"

##  [1] "modelName"                "protein_Id"              
##  [3] "meanAbundanceImp_group_1" "meanAbundanceImp_group_2"
##  [5] "diff"                     "group_1_name"            
##  [7] "group_2_name"             "contrast"                
##  [9] "avgAbd"                   "indic"                   
## [11] "nrMeasured_group_1"       "nrMeasured_group_2"      
## [13] "df"                       "sigma"                   
## [15] "std.error"                "statistic"               
## [17] "p.value"                  "conf.low"                
## [19] "conf.high"                "FDR"

Finally we are merging the results and give priority to the results where we do not have missing values in one group.

merged <- prolfqua::merge_contrasts_results(prefer = contr,add = mC)$merged
plotter <- merged$get_Plotter()
tmp <- plotter$volcano()
tmp$FDR

Volcano plots for the two contrasts with missing value imputation from the group_average model.

Look at proteins which could not be fitted using the linear model, if any.

merged <- prolfqua::merge_contrasts_results(prefer = contr,add = mC)

moreProt <- transformed$get_copy()
moreProt$data <- moreProt$data |> dplyr::filter(protein_Id %in% merged$more$contrast_result$protein_Id)
moreProt$get_Plotter()$raster()

# here we do not get anything because there is nothing imputed!

GSEA Analyis

We can rank the proteins based on the log2FC or the t-statistic and subject them them to gene set enrichment analysis (see GSEA).

This example will run only if the following packages are installed on you machine:

clusterProfiler (Bioconductor)
org.Sc.sgd.db (Bioconductor)
prora (github.com/protviz/prora)

#evalAll <- require("clusterProfiler") & require("org.Sc.sgd.db") & require("prora")
evalAll <- require("clusterProfiler") & require("org.Sc.sgd.db2") & require("prora")

library(clusterProfiler)
library(org.Sc.sgd.db)

bb <- prolfqua::get_UniprotID_from_fasta_header(merged$merged$get_contrasts(),
                                             idcolumn = "protein_Id")
bb <- prora::map_ids_uniprot(bb)
ranklist <- bb$statistic
names(ranklist) <- bb$P_ENTREZGENEID
res <- clusterProfiler::gseGO(
  sort(ranklist, decreasing = TRUE),
  OrgDb = org.Sc.sgd.db,
  ont = "ALL")

ridgeplot( res )

dotplot(res , showCategory = 30)

enrichplot::upsetplot(res)

The prolfqua package is described in (Wolski et al. 2022).

Session Info

sessionInfo()

## R version 4.3.2 (2023-10-31)
## Platform: aarch64-apple-darwin20 (64-bit)
## Running under: macOS Sonoma 14.5
## 
## Matrix products: default
## BLAS:   /Library/Frameworks/R.framework/Versions/4.3-arm64/Resources/lib/libRblas.0.dylib 
## LAPACK: /Library/Frameworks/R.framework/Versions/4.3-arm64/Resources/lib/libRlapack.dylib;  LAPACK version 3.11.0
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## time zone: Europe/Zurich
## tzcode source: internal
## 
## attached base packages:
## [1] stats     graphics  grDevices utils     datasets  methods   base     
## 
## other attached packages:
## [1] prolfqua_1.2.5
## 
## loaded via a namespace (and not attached):
##  [1] gtable_0.3.5       xfun_0.43          bslib_0.7.0        ggplot2_3.5.1     
##  [5] htmlwidgets_1.6.4  ggrepel_0.9.5      vctrs_0.6.5        tools_4.3.2       
##  [9] crosstalk_1.2.1    generics_0.1.3     tibble_3.2.1       fansi_1.0.6       
## [13] highr_0.10         pkgconfig_2.0.3    pheatmap_1.0.12    KernSmooth_2.23-22
## [17] data.table_1.15.4  RColorBrewer_1.1-3 desc_1.4.3         lifecycle_1.0.4   
## [21] compiler_4.3.2     farver_2.1.1       stringr_1.5.1      textshaping_0.3.7 
## [25] progress_1.2.3     munsell_0.5.1      httpuv_1.6.15      htmltools_0.5.8.1 
## [29] sass_0.4.9         yaml_2.3.8         lazyeval_0.2.2     plotly_4.10.4     
## [33] later_1.3.2        pillar_1.9.0       pkgdown_2.0.9      crayon_1.5.2      
## [37] jquerylib_0.1.4    tidyr_1.3.1        MASS_7.3-60.0.1    cachem_1.0.8      
## [41] mime_0.12          tidyselect_1.2.1   conflicted_1.2.0   digest_0.6.35     
## [45] stringi_1.8.3      dplyr_1.1.4        purrr_1.0.2        labeling_0.4.3    
## [49] forcats_1.0.0      fastmap_1.1.1      grid_4.3.2         colorspace_2.1-0  
## [53] cli_3.6.2          magrittr_2.0.3     utf8_1.2.4         withr_3.0.0       
## [57] promises_1.3.0     prettyunits_1.2.0  scales_1.3.0       rmarkdown_2.26    
## [61] httr_1.4.7         gridExtra_2.3      ragg_1.3.0         hms_1.1.3         
## [65] shiny_1.8.1.1      memoise_2.0.1      evaluate_0.23      knitr_1.46        
## [69] viridisLite_0.4.2  rlang_1.1.3        Rcpp_1.0.12        xtable_1.8-4      
## [73] glue_1.7.0         jsonlite_1.8.8     R6_2.5.1           systemfonts_1.0.6 
## [77] fs_1.6.4

References

Wolski, Witold E., Paolo Nanni, Jonas Grossmann, Maria d’Errico, Ralph Schlapbach, and Christian Panse. 2022. “Prolfqua: A Comprehensive R-package for Proteomics Differential Expression Analysis.” bioRxiv. https://doi.org/10.1101/2022.06.07.494524.

Witold E. Wolski

2024-05-30