README.md
In crossmeta: Cross Platform Meta-Analysis of Microarray Data

Cross-Platform Meta-Analyis

crossmeta streamlines the cross-platform meta-analysis of microarray data. For the analysis, you will need a list of Affymetrix, Illumina, and/or Agilent GSE numbers from GEO. All 21 species in the current homologene build are supported. See vignette for detailed usage.

crossmeta is available through Bioconductor. To install the latest release from github:

install.packages('remotes')
remotes::install_github('alexvpickering/crossmeta')

# studies from GEO
gse_names  <- c("GSE9601", "GSE15069")

# get raw data for specified studies
get_raw(gse_names)

# load and annotate raw data
esets <- load_raw(gse_names)

# perform differential expression analysis
anals <- diff_expr(esets)

# perform meta-analysis
es <- es_meta(anals)

A high quality meta-analysis is achieved by addressing the key issues in conducting a meta-analysis of microarray data (1):

Uses raw data

Different labs process their raw data differently. These differences in data processing may lead to flawed conclusions upon meta-analysis.
crossmeta starts with raw data, and uses a consistent processing pipeline for all studies.

Maps probes to human genes

Related genes from different species are often referenced with different symbols. These differences can make it challenging to compare similar microarray experiments in diverse species.
crossmeta maps probes to human gene symbols using homology relationships established by HomoloGene.

Resolves many-to-many mappings

Single probes can measure multiple genes. To incorporate these measurements appropriately, they are replaced with a new record for each gene.
Multiple probes can measure the same gene. Averaging these measurements is inappropriate because measurement scales will vary with probe affinity. crossmeta selects the measurement with the highest inter-quartile range as it is the least likely to occur by chance.

Models nuisance variables

In addition to variables of interest, there are sources of signal due to factors that are unknown, unmeasured, or too complicated to capture through simple models. These factors can either hide true effects or introduce spurious ones.
crossmeta discovers and accounts for these nuissance variables using surrogate variable analysis.

Simplifies model specification

Correctly specifying a model and contrast matrix can be challenging.
crossmeta uses an attractive user interface that allows you to simply select the samples you want to compare.
Paired samples (eg. the same subject before and after treatment) can also be selected using the same interface.

Meta-analyzes genes with missing data

Differences in microarray platforms often lead to thousands of genes that are not measured in all studies. Most existing meta-analysis software requires that these genes with missing data are discarded.
crossmeta extends the effect size meta-analysis method in GeneMeta to allow for genes that were not measured in all studies. Keep your data, find more insights.

(1) Ramasamy A, Mondry A, Holmes CC, Altman DG (2008) Key Issues in Conducting a Meta-Analysis of Gene Expression Microarray Datasets. PLoS Med 5(9): e184. doi:10.1371/journal.pmed.0050184