In fmicompbio/mutscan: Preprocessing and Analysis of Deep Mutational Scanning Data

knitr::opts_chunk$set(echo = TRUE)

suppressPackageStartupMessages({
    library(SummarizedExperiment)
    library(mutscan)
    library(DT)
    library(dplyr)
})

Data object

The input data use in this report is contained in the following SummarizedExperiment object:

params$se

Sample annotation and filtering summary

The table below summarizes the available sample information, including annotations provided by the data analyst as well as filtering information collected by mutscan.

tmpdf <- as.data.frame(colData(params$se))
extradf <- do.call(
    dplyr::bind_rows, 
    lapply(names(metadata(params$se)$parameters), function(nm) {
        data.frame(Name = nm, 
                   lapply(metadata(params$se)$parameters[[nm]], 
                          function(x) paste(x, collapse = "; ")))
    })
)
DT::datatable(dplyr::full_join(tmpdf, extradf, by = "Name"),
              extensions = "FixedColumns",
              rownames = FALSE,
              options = list(
                  paging = TRUE, searching = TRUE, info = FALSE,
                  pageLength = 20, 
                  sort = TRUE, scrollX = TRUE, fixedColumns = list(leftColumns = 1)
              )) %>%
    formatStyle("Name", "vertical-align" = "center")

The plots below focus on the filtering information, and show the number of reads remaining after each filtering step performed by mutscan, as well as the fraction of reads that are filtered out by each step.

tryCatch({mutscan::plotFiltering(params$se, valueType = "reads", onlyActiveFilters = TRUE, 
                                 displayNumbers = TRUE, plotType = "remaining", facetBy = "sample")},
         error = function(e) message("Couldn't generate filtering plots."))

tryCatch({mutscan::plotFiltering(params$se, valueType = "fractions", onlyActiveFilters = TRUE, 
                                 displayNumbers = TRUE, plotType = "filtered", facetBy = "step")},
         error = function(e) message("Couldn't generate filtering plots."))

Pairs plot

Next, we show a pairs plot, displaying the correlation between each pair of samples.

tryCatch({mutscan::plotPairs(params$se)},
         error = function(e) message("Couldn't generate pairs plot."))

Total count plot

The plot below shows the total count for each sample, across all features in the SummarizedExperiment object.

tryCatch({mutscan::plotTotals(params$se, selAssay = "counts")},
         error = function(e) message("Couldn't generate total count plot."))

Distribution plots

The next group of plots displays the distribution of counts for each sample (across all features in the SummarizedExperiment object). The same information is displayed as a 'knee plot', where the counts for each sample are arranged in decreasing order, and as a density plot. Note that both plots display the values on a log scale; hence any zero values are not shown.

tryCatch({mutscan::plotDistributions(params$se, selAssay = "counts", pseudocount = 1,
                                     groupBy = NULL, plotType = "knee", facet = FALSE)},
         error = function(e) message("Couldn't generate distribution plot."))

tryCatch({mutscan::plotDistributions(params$se, selAssay = "counts", pseudocount = 1, 
                                     groupBy = NULL, plotType = "density", facet = FALSE)},
         error = function(e) message("Couldn't generate distribution plot."))