imprints_IS: imprints_IS
In mgerault/mineCETSAapp: A shiny app for IMPRINTS.CETSA package

imprints_IS

R Documentation

imprints_IS

Description

Function to categorize protein according to their Expression/Stability change based on their I-score and their combined p-value on their two biggest fold changes.

Usage

imprints_IS(
  data,
  data_diff = NULL,
  ctrl,
  valid_val = NULL,
  IS_cutoff = 1.5,
  fixed_score_cutoff = FALSE,
  FDR = 0.01,
  pv_method = c("all", "top2"),
  adj_pv_method = "BH",
  comb_pv_method = c("fisher", "george", "edgington"),
  curvature = 0.05,
  FDR_category = 0.1,
  format_category = c("9", "4"),
  folder_name = "Hits_analysis",
  peptide_count_col = "sumUniPeps",
  species = "Homo Sapiens"
)

Arguments

`data`	The input data set from which categorization is performed on and hitlist is produced from.
`data_diff`	The output from imprints_caldiff; can be NULL and if so will compute it; can also be a path to the data.
`ctrl`	The name of the control.
`valid_val`	The percentage of non-missing values you want per treatment. If less, score will be set to NA, i.e. the protein will not be a hit
`IS_cutoff`	The I-score cutoff. Default is 1.5.
`fixed_score_cutoff`	Logical to tell if you want to use a fixed cutoff for the I-score. If TRUE, the value IS_cutoff will directly be used as the cutoff and for all treatments. If FALSE, the I-score cutoff will be calculated as the value selected for IS_cutoff plus the median of the I-scores of the proteins which have a p-value lower than the median of all p-values for a given treatment. Default is FALSE.
`FDR`	The FDR used for the BH corrected combined p-value
`pv_method`	The method to compute the p-values. If top2, then only the p-values from the two greatest fold-changes will be computed and combined; if all, will take all fold-changes. Default is all.
`adj_pv_method`	see `p.adjust`; default is BH
`comb_pv_method`	The method used to combine p-values. Either george, fisher or edgington. Default is fisher; see Details.
`curvature`	The curvature used for the curve on the volcano plot
`FDR_category`	The FDR used for the BH corrected p-value at 37°C used in order to categorize the hits
`format_category`	Choose between 9 or 4, indicating how many categories to segregate the hits; default value is 9. If 9 is selected, the 9 categories will be: NN, CN+, CN-, NC+, NC-, CC++, CC+-, CC-+, CC–. If 4, then it will be: NN, CN, NC, CC. The sign of + or - after N or C is determined by the sign of the fold-change at 37°C and the signof the I-score, respectively.
`folder_name`	The name of the folder in which you want to save the results.
`peptide_count_col`	The name of the column that contain the unique peptide count. If it is sumUniPeps, don't bother with this parameter.
`species`	The species on which you did the experiment (not necessary if already present in your datas). Default is 'Homo Sapiens'.

Details

George's method correspond to the sum of the logit of the p-values, Fisher's to the sum of the log of the p-values and Edgington's is the sum of the p-values. Edgington's method is the most stringent and is particularly sensitive with higher p-values whereas Fisher's method is the less stringent as it is mostly sensitive to low p-values. George's method is a compromise between the two methods. For more details read https://doi.org/10.48550/arXiv.1707.06897.