iDEA.BMA: Bayesian model averaging (BMA) approach to aggregate DE...

In xzhoulab/iDEA: Integrative Differential Expression and Gene Set Enrichment Analysis

Bayesian model averaging (BMA) approach to aggregate DE evidence for any given genes across all available gene sets without the requirement of pre-selecting a gene set. Specifically, for the given gene, we denote its posterior inclusion probability (PIP) obtained using the gene set k as PIP_k. The corresponding Bayes factor quantifying its DE evidence based on the gene set k is BF_k=PIP_k/(1-PIP_k). With equal prior weights on different gene sets, the average Bayes factor quantifying its DE evidence based on all K gene sets is thus ABF, which can be converted back to a posterior inclusion probability as PIP=ABF/(1+ABF).

Description

Bayesian model averaging (BMA) approach to aggregate DE evidence for any given genes across all available gene sets without the requirement of pre-selecting a gene set. Specifically, for the given gene, we denote its posterior inclusion probability (PIP) obtained using the gene set k as PIP_k. The corresponding Bayes factor quantifying its DE evidence based on the gene set k is BF_k=PIP_k/(1-PIP_k). With equal prior weights on different gene sets, the average Bayes factor quantifying its DE evidence based on all K gene sets is thus ABF, which can be converted back to a posterior inclusion probability as PIP=ABF/(1+ABF).

Usage

iDEA.BMA(object)

Arguments

object	iDEA object
...	Ignored

Value

Returns a iDEA object with posterior inclusion probability averaged by all gene sets results in object@BMA_pip.

xzhoulab/iDEA documentation built on Oct. 8, 2022, 8:54 a.m.