MAPE2.0: Perform the Meta-Analysis for Pathway Enrichment

Description Usage Arguments Details Value Author(s) Examples

Description

This is the first major function in the MetaPath2.0 package which combines the Meta-analysis for Pathway Enrichment (MAPE) methods introduced by Shen and Tseng (2010) and the Comparative Pathway Integrator (CPI) method introduced by Fang and Tseng (2016). The default function is CPI which performs MAPE_P (integrating multiple studies at pathway level) with Adaptively Weighted Fisher's method as Meta-analysis statistics. The other function, MAPE, automatically performs MAPE_G (integrating multiple studies at gene level), MAPE_P and MAPE_I (a hybrid method integrating MAEP_G and MAPE_P methods).

Usage

 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
MAPE2.0(
  arraydata = NULL,
  clinical.data = NULL,
  label = NULL,
  pmtx = NULL,
  pathway = c(Biocarta.genesets, GOBP.genesets, GOCC.genesets, GOMF.genesets,
    KEGG.genesets, Reactome.genesets),
  data.type = c("continuous", "discrete"),
  mixed = NULL,
  mix.type = NULL,
  covariate = NULL,
  ref.level,
  paired,
  ind.method,
  select.group,
  tail = "abs",
  resp.type = c("twoclass", "multiclass", "continuous", "survival"),
  method = c("CPI", "MAPE"),
  enrichment = c("KS", "Fisher's exact"),
  DEgene.number = 200,
  stat = c("Fisher", "maxP", "minP", "rth", "AW Fisher"),
  rth.value = NULL,
  permute = F,
  nperm = 500,
  size.min = 15,
  size.max = 500,
  qvalue.cal = c("estimate", "permutation")
)

Arguments

arraydata

The arraydata is a list of microarray data sets. Each microarray data set can be either an ExpressionSet or a list. If the microarray data set is a list, then it includes five elements as follows: 1)x-exprs data 2)y- the phenotype of interests 3)z- censoring.status if applicable. 1 stands for the event occurred and 0 stands for censored. 4)geneid 5)samplename If the microarray data set is in an ExpressionSet format, the users need to 1) store the phenotype of interests in the slot 'label'. 2) store the censor data is the slot 'censoring.status' if applicable

clinical.data

Clinical data files

label

Label selected from clinical.data

pmtx

Option for uploading p-value matrix.

pathway

The pathway databases

data.type

It is one of the two values: 'continuous','discrete'.

resp.type

The phenotype of interest.It is one of the four values: 'twoclass','multiclass', 'continuous','survival'.

method

The method for overall analysis. It is one of the two values: 'CPI','MAPE'.

enrichment

The method for pathway enrichment analysis. It is one of the two values: 'KS', 'Fisher's exact'.

DEgene.number

For Fisher's exact method, the number of differentially expressed genes identified from each studies.

stat

The meta-analysis statistic to be used to combine multiple studies. For MAPE, it is one of the five values: 'minP','maxP','rth','Fisher','AW Fisher'.For CPI, AW Fisher's method is the only option.

rth.value

The value of the rth statistic if the meta-anlaysis statistic is 'rth'. For example,rth.value = 2.

permute

Whether to use permutation to calculate p-values. By default, it is false.

nperm

Number of permutations to be performed.

size.min

The minimum size of pathways to be considered. The default value is 15.

qvalue.cal

The method to calculate the q-values if permute is true. The default method is to calcuate the q-values based on the permutation method. If qvalue.cal='estimate', the q-values were estimated based on the Storey's method.

size

max The maximum size of pathways to be considered. The default value is 500.

Details

For MAPE, in the simulation and real data analyses in the paper, MAPE_G and MAPE_P have complementary advantages and detection power depending on the data structure. In general, the integrative form of MAPE_I is recommended to use. In the case that MAPE_G (or MAPE_P) detects almost none pathway, the integrative MAPE_I does not improve performance and MAPE_P (or MAPE_G) should be used.

Value

The qvalue and pvalue of each pathway including Meta qvalue and Meta pvalue.

Author(s)

Kui Shen, Xiangrui Zeng and George Tseng.

Examples

1
2
3
4
5
6
7
8
data(Psychiatry_disease_filtered)
data(pathways)
CPI_result = MAPE2.0(arraydata = Psychiatry_diseases$expr, clinical.data = Psychiatry_diseases$clinical,
                    label = "response",pmtx = NULL,pathway = c(Biocarta.genesets,GOBP.genesets,GOCC.genesets,GOMF.genesets,
                    KEGG.genesets,Reactome.genesets),data.type ="continuous", resp.type = "twoclass",method = "CPI",
                    ind.method = rep("limma",length(Psychiatry_diseases$expr)),paired = rep(FALSE,length(Psychiatry_diseases$expr)),
                    select.group = c("CASE","CTRL"),ref.level ="CTRL",tail="abs",
                    enrichment = "Fisher's exact", DEgene.number = 400,stat = "AW Fisher")

metaOmics/MetaPath documentation built on June 15, 2020, 10:22 a.m.