EnrichNet Network-based gene set enrichment analysis

2015-01-13

EnrichNet Network-based gene set enrichment analysis

EnrichNet: network-based gene set enrichment analysis
http://www.enrichnet.org/

Luxembourg/France/Nottingham/CNIO paper
Bioinformatics journal 2012


EnrichNet Network-based gene set enrichment analysis


q
The analysis of functional genomics data from high-throughput
experiments often involves the assessment of potential functional
associations between a gene or protein set of interest, e.g.
differentially expressed genes in a microarray study and known
gene/protein sets representing cellular processes and pathways.


?I keep forgetting what it means if something is parametric or non-parametric?

q
EnrichNet,
a new integrative enrichment analysis method.


q
This combined network analysis and
visualization enables a direct molecular interpretation of how a user-
de?ned set of genes/proteins is related to a gene/protein set of known
function.

q
To further facilitate the analysis, a complete
implementation of the integrative approach is made freely available
as a public web application with an exposed programmatic API
(www.enrichnet.org)


q
The only required input is a list of 10 or more human gene
or protein identi?ers and the selection of a database of interest (KEGG
(Kanehisa et al., 2006), BioCarta (Nishimura, 2001), WikiPathways (Pico
et al., 2008), Reactome (Joshi-Tope et al., 2005), PID (Schaefer et al., 2009),
InterPro (Apweiler et al., 2001) or GO (Ashburner et al., 2002), from which
reference gene/protein sets will be extracted.

q
A random walk on a graph is a stochastic process
modelling the iterative transition of an imaginary particle from a seed node. . .

q
in spite of its name, RWR is a
deterministic procedure modelling a random walk via matrix computations

q
Although the
entire procedure is more computationally expensive than a classical over-
representation analysis, this does not result in signi?cant limitations for
practical use, since an analysis takes only a few minutes for most pathway
databases, and the web-interface optionally provides an e-mail noti?cation
for completed tasks

q
Importantly, EnrichNet and ORA methods are
designed speci?cally for the analysis of gene/protein lists that are not
accompanied by any additional expression or activity measurements, and
microarray data are used only for validation purposes.

q
In all
cases, EnrichNet provides higher enrichment scores than the ORA
approach and its P-value estimates are either lower or below the
detection limit (0.001) for both methods.

q
The high Xd-score
for this gene set pair (0.80, the signi?cance threshold is 0.45)
points to a functional association via multiple connecting molecular
interactions, which is con?rmed by the visualization

q
illustrate the utility of the approach for
identifying novel functional associations between gene/protein sets