deg.specific: Predict the patient-specific DEGs using bi-clustering...

In DEComplexDisease: A tool for differential expression analysis and DEGs based investigation to complex diseases by bi-clustering analysis

Description

This function is use the output of bi.deg as input to predict the differentially expressed genes (DEGs) for each patient by cross-validation of multiple patients.

Usage

deg.specific(deg, test.patients = NULL, min.genes = 50, min.patients = 5,
  overlap = 0.85, cores = 1)

Arguments

deg	a 'deg' object. This should be the output of bi.deg
test.patients	the patients to test. Only the patients in 'test.patients' are used as seed for bi-clustering analysis
min.genes	the minimum number of genes
min.patients	the minimum number of patients. It includes the patients as seeds (see details).
overlap	the minimum similarity for selected DEGs from two or more patients
cores	the thread number

Details

The DEGs from bi.deg are mixed with noises, e.g. the DEGs not associated with disease. This is especially true when the differential expression analysis tests are done using one variable again references. The assumption behind his analysis is that the disease associated DEGs will be observed in other patients. This function implements a bi-clustering algorithm to find the DEGs shared by 'min.patients' in the binary DEG matrix. In this process, each patient is used as seed and its DEGs are gradually excluded to find if there is a DEG list which is observed in 'min.patients' when the similarity is greater is 'overlap'.

'test.patients' option is used to find the cross-validated DEGs for some interest patients. Otherwise, all the patients will be used as seeds. 'overlap' is the threshold to determine the minimum similarity between neighbor and seed patients.

Value

A 'deg.specific' or 'deg.specific.test' object. It has a key of 'decd.input', which stores the binary DEG matrix, genes, patients and used parameter setting. Other keys are the patients IDs and they store the cross-validated DEGs.

Author(s)

Guofeng Meng

Examples

# the DEGs has at least 100 genes and validated by 5 other patients
deg.spc <- deg.specific(deg, min.genes=50, min.patients=5, overlap=0.85)

DEComplexDisease documentation built on Nov. 8, 2020, 6:42 p.m.