Skip to content
This repository has been archived by the owner on Feb 11, 2020. It is now read-only.

llrs/GSEAdv

Repository files navigation

GSEAdv

Travis build status AppVeyor Build Status Coverage status lifecycle Project Status: Abandoned – Initial development has started, but there has not yet been a stable, usable release; the project has been abandoned and the author(s) do not intend on continuing development.

The goal of GSEAdv is to provide methods to work with gene sets collections.

This package is abandonded as a new package is being developed to work with gene sets under Bioconductor. My own efforts are now in BaseSet

GSEAdv is based on the relationship between genes and gene sets under this schema: Schema of gene sets

And provides methods to understand the relationships between each property of the schema and as a whole.

Installation

It is an R package you will be able to install it from the Bioconductor project with:

## install.packages("BiocManager") 
BiocManager::install("GSEAdv")

You can install this version of GSEAdv with:

## install.packages("devtools")
devtools::install_github("llrs/GSEAdv")

How does it work?

It is simple, load the package and learn from your data!

# Load some data
library("GSEAdv")
fl <- system.file("extdata", "Broad.xml", package = "GSEABase")
gss <- getBroadSets(fl)
gss
## GeneSetCollection
##   names: chr5q23, chr16q24 (2 total)
##   unique identifiers: ZNF474, CCDC100, ..., TRAPPC2L (215 total)
##   types in collection:
##     geneIdType: SymbolIdentifier (1 total)
##     collectionType: BroadCollection (1 total)
summary(gss)
## Genes: 215
##  Gene in more pathways: 1 pathways
##  h-index: 0 genes with at least 0 pathways.
## Pathways: 2
##  Biggest pathway: 129 genes
##  h-index: 1 pathways with at least 1 genes.
## All genes in a single gene set.

Which tells us that each gene in the GeneSetCollection is only on one gene set.

We can try with a bigger dataset, one derived from human genes pathways in KEGG:

summary(genesKegg)
## Genes: 5869
##  Gene in more pathways: 51 pathways
##  h-index: 0 genes with at least 0 pathways.
## Pathways: 228
##  Biggest pathway: 1130 genes
##  h-index: 15 pathways with at least 15 genes.
## IC(genesPerPathway): 6.65 ( 0.96 of the maximum)
## IC(pathwaysPerGene): 2.47 ( 0.48 of the maximum)

Knowing that it has so much pathways and genes we can learn how do they relate. The number of genes per pathway in the collection is:

gpp <- genesPerPathway(genesKegg)
plot(table(gpp))

Distribution of the number of genes per gene set. We can see that most gene sets have low number of genes but one has 1130 genes in a single gene set (It is the gene set 01100). The genes might be associated too with many gene sets, it is so extreme? Let's see:

ppg <- pathwaysPerGene(genesKegg)
plot(table(ppg))

Distribution of the number of gene sets per gene Not so extreme, one gene (5594) appears in 51 gene sets.

To see which gene sets are included in other gene sets we can use nested:

nested(genesKegg)[1:10, 80:90]
##       00970 00980 00982 00983 01040 01100 02010 03008 03010 03013 03015
## 00010     0     0     0     0     0     0     0     0     0     0     0
## 00020     0     0     0     0     0     1     0     0     0     0     0
## 00030     0     0     0     0     0     0     0     0     0     0     0
## 00040     0     0     0     0     0     0     0     0     0     0     0
## 00051     0     0     0     0     0     0     0     0     0     0     0
## 00052     0     0     0     0     0     0     0     0     0     0     0
## 00053     0     0     0     0     0     0     0     0     0     0     0
## 00061     0     0     0     0     0     1     0     0     0     0     0
## 00062     0     0     0     0     0     1     0     0     0     0     0
## 00071     0     0     0     0     0     0     0     0     0     0     0

As expected the pathway with more than 1100 genes has other pathways inside it.

You can see the vignettes for more examples.

Who will use this repo or project?

It is intended for bioinformaticians, both people interested in comparing databases and people developing analysis using the information provided by GSEAdv.

What is the goal of this project?

The goal of this project is to be able to understand the gene sets collections available.

What can be GSEAdv used for?

  • Measure properties of the genes and pathways given a GeneSetCollection The number of genes by pathway, the probability of having x genes in more than y pathways...
  • Compare pathway database: By comparing the differences between them.
  • Select the gene set collection of interest: By testing their properties.
  • Create GeneSetCollections with certain properties Create a GeneSetCollection were the collections follow certain distributions.

Contributing

Please read how to contribute for details on the code of conduct, and the process for submitting pull requests.

You can also look at the tests and add more tests to increase the quality of the package.

Acknowledgments

The ideas of this package were developed after a colleague asked a question in a poster presentation of my other package BioCor. To know the whole history you can read this blogpost.

Releases

No releases published

Packages

No packages published

Languages