panelOptimizer: A function that open a shiny app to optimize panel size while...

In gmelloni/PrecisionTrialDesigner: A Tool to Analyze and Design NGS Based Custom Gene Panels

A function that open a shiny app to optimize panel size while retaining the majority of mutations

Description

Given a CancerPanel object, it automatically reads the genes and panel information and allows a custom subsetting of the panel to retain the largest amount of mutations while saving genomic space

Usage

panelOptimizer(object)

Arguments

object

A CancerPanel object filled with genomic data.

Details

This function reads the panel objects and retrieves all the genes requested for SNV and full sequence. It also collects all the mutation data and the tumor types available and performs an analysis on the position of the mutations on the canonical protein sequence for every gene.

At function call, a shiny app with four tabs is opened. Select a gene and the tumor types to use and click Run.

On the first and second tabs, the user is guided in the choice of most appropriate regions using an in-house Bioconductor package called LowMACA. LowMACA creates a null model where all the mutations on each gene are randomly permutated along the sequence. Every position that exceeds the threshold of 95% confidence interval is considered not random and represents an hotspot. The user can also decide to use a bandwidth in this calculation and apply a Gaussian density to the distribution of mutations along the sequence. Alterations that are closed to each other in the sequence will be aggregated to form significant regions. If the LowMACA analysis succeded, a table will appears under the plot in the first tab and on the second tab a list of significant positions is also shown. Click on Store LowMACA yellow button and all the regions identified by the algorithm will be stored in the fourth tab (Optimize Panel).

If you want to select your own custom regions, go to the third tab (Manual Selection). You can click on the red dots and retrieve information on specific mutations. If you drag a region with the mouse, a table will appear below, with the same information as the LowMACA analysis. Click on Store yellow button to keep the region selected and move it to the fourth tab.

For a new analysis on a different gene, just select the gene and click Run again.

When satisfied, just click on Close and save or simply close the browser page. All the regions selected will be merged and returned in standard output.

Value

A list of three elements after closing the shiny session.

1. regions a data.frame with the regions selected, the percentage of space occupied and percentage of mutations captured

2. mergedRegions merged protein regions by gene from user section

3. panel if no regions were selected, it reports the original panel, otherwise it substitutes the gene requested in full sequence with the regions in mergedRegions

Author(s)

Giorgio Melloni, Alessandro Guida

See Also

entropy lmPlot

Examples

## Only run this example in interactive R sessions
if (interactive()) {
  # Load example CancerPanel object
  data(cpObj)
  # Optimize the space on the shiny app. 
  # All changes mad on the app will be saved 
  newpanel <- panelOptimizer(cpObj)
  # If some changes have been made, recreate a new CancerPanel object
  if(!is.null(newpanel$regions)){
    cpObjOptimized <- newCancerPanel(newpanel$panel)
    # Fill the object with the same data of the non optimized panel
    cpObjOptimized <- getAlterations(cpObjOptimized , repos=cpData(cpObj))
    # Subset alterations on the new panel directives
    cpObjOptimized <- subsetAlterations(cpObjOptimized)
  }
}

gmelloni/PrecisionTrialDesigner documentation built on March 3, 2023, 6:10 a.m.