top_table

Share:

Description

Get IDs that pass two filters, p-value and effect-size. This top_table will make a significant list that is less than p-value and greater than effect-size. Effect-size are calculated by obtained power level. This function requires four parameters. ex) top_table(input_data,pvalue=0.05,power_desired=0.90,method='QPF')

Usage

1
2
top_table(input_list, pvalue = 0.05, power_desired = 0.9, method = "QPF",
  FC_threshold = 2)

Arguments

input_list

The list should be produced by 'get_statistics_from_file' or 'get_statistics_from_dataFrame' function. See get_statistics_from_file and get_statistics_from_dataFrame for more information. It consists of the following items:

$data_table A data frame that have statistics for each IDs
$min_rep Common number of replicates in your group information.
$max_rep Maximum number of replicates in your group information.
$nt The number of total experiments in your expression profile.
$ng The number of groups in your group information.
pvalue

p-value should be ranged between 0 to 1. default is 0.05.

power_desired

Give the statistical power you desired for output significant list

method

Choose statistics method you want to use for making significant list

"QPF" combination of Quasi-Poisson and Cohen's f. Default.
"QPF2" combination of Quasi-Poisson and Cohen's f2.
"QPFC" combination of Quasi-Poisson and Fold change.
"NBW" combination of Negative Binomial and Cohen's w.
"NBF2" combination of Negative Binomial and Cohen's f2.
"NBFC" combination of Negative Binomial and Fold change.
"NORF" combination of ANOVA with normal distribution and Cohen's f.
"NORFC" combination of ANOVA with normal distribution and Fold change.
FC_threshold

Fold change you want to use. Default is 2.

Value

A list containing the follow items and a scatter plot that x-axis is effect size and y-axis is probability. Vertical line the plot is minimum effect size and horizontal line is maximum probability threshold. Red dots means insignificant, while blue dots are significant.

top_table a data frame that have calculated statistics for top table IDs
minimum_effect_size Minimum effect size threshold
selected_effect_size_filter The selected effect size filter
minimum_power Minimum statistical power in the top_table
selected_model The selected probability model for calculating p-value
alpha Maximum adjusted p-value
method_pvalue_adjustment The selected method for p-value adjustment
num_group The number of groups used for generating the top_table
common_replicates The number of common replicates.
num_columns The number of columns (samples or experiments)
top_table's elements
Cohens_W Cohen's w
Cohens_F Cohen's f
Cohens_F2 Cohen's f2
Max_FC Maximum fold change among all the possible group pairs
QP_Pval_adjusted Adjusted p-value from GLM quasi-Poisson
NB_Pval_adjusted Adjusted p-value from GLM negative binomial
Normal_Pval_adjusted Adjusted p-value from Normal ANOVA

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
library(selfea)

## Test selfea for single protein expression
values <- c(6,8,10,29,26,22)
groups <- c("U200","U200","U200","U600","U600","U600")
experiments <- c("exp1","exp2","exp3","exp4","exp5","exp6")

df_expr <- data.frame(ID="Protein_1",exp1=6,exp2=8,exp3=10,exp4=29,exp5=26,exp6=22)
df_group <- data.frame(Col_Name=experiments,Group=groups)
list_result <- get_statistics_from_dataFrame(df_expr,df_group)
top_table(list_result)

## For this example we will import Gregori data
## Josep Gregori, Laura Villareal, Alex Sanchez, Jose Baselga, Josep Villanueva (2013).
## An Effect Size Filter Improves the Reproducibility
## in Spectral Counting-based Comparative Proteomics.
## Journal of Proteomics, DOI http://dx.doi.org/10.1016/j.jprot.2013.05.030')

## Description:
## Each sample consists in 500ng of standard yeast lisate spiked with
## 100, 200, 400 and 600fm of a mix of 48 equimolar human proteins (UPS1, Sigma-Aldrich).
## The dataset contains a different number of technical replimessagees of each sample

## import Gregori data
data(example_data1)
df_contrast <- example_data
df_group <- example_group

## Get statistics through 'get_statistics_from_dataFrame' function
list_result <- get_statistics_from_dataFrame(df_contrast,df_group)

## Get significant features (alpha >= 0.05 and power >= 0.90)
significant_qpf <- top_table(list_result,pvalue=0.05,power_desired=0.90,method='QPF')