selfea (version 1.0.1)

get_statistics_from_dataFrame: get_statistics_from_dataFrame

Description

This function computes Cohen's f, f2 and w, adjusted p-value from GLM quasi-Poisson, negative binomial and Normal distribution.

Usage

get_statistics_from_dataFrame(df_contrast, df_group, padj = "fdr")

Arguments

df_contrast

A data frame that consists of 'ID' column and expression profile (columns after 'ID' column). 'ID' column should be unique. Column names after 'ID' column should be unique. Only positive numbers are allowed in expression data. Here is an example.

ID	Y500U100_001	Y500U100_002	Y500U200_001
Y500U200_002	YKL060C	151	195
188	184	YDR155C	154
244	237	232	YOL086C
64	89	128	109
YJR104C	161	155	158
172	YGR192C	157	161
173	175	YLR150W	96
109	113	115	YPL037C
23	28	27	27
YNL007C	53	58	64
63	YBR072W	52	53
54	44	YDR418W_1	76
53	62	74	ID

df_group

A data frame that consists of 'Col_Name' and 'Group' columns This parameter is to match experiment groups to expression profiles of df_contrast. 'Col_Name' should be corresponding to column names of expression profile of df_contrast. 'Group' columns have experiment informaion of columns in expression profile of df_contrast. Here is an example. See the example of df_contrast together.

Col_Name

Group

Y500U100_001

U100

Y500U100_002

U100

Y500U200_001

U200

Y500U200_002

U200

padj

Choose one of these c("holm", "hochberg", "hommel", "bonferroni", "BH", "BY", "fdr", "none"). "fdr" is default option. The option is same to p.adjust.

Value

A list that 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.

$method_pvalue_adjustment

The selected method for p-value adjustment

data_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

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')