are indicated. One would need to compare these 6-mer
sequences to databases already previously described to see if
there is overlap with a known sequence. Otherwise, promoter
deletion experiments in the lab may show necessity of this
sequence for expression, and serve as a starting point for further experimental analyses.
3.7 Functional
Classification
Functional classification of gene lists is one of the basic methods in
bioinformatics for making sense of sometimes rather large gene lists
that arise from gene expression profiling experiments. Typically,
one might look at individual genes in such lists and “see if they fit
biologically,” but one might also like to have an overview of broad
functional categories that change in response to a given stimulus or
due to a specific mutation. One of the very useful large initiatives of
the past decade was the development of a Gene Ontology (GO) for
the “unification of biology” [50]. Basically, this system is based on a
set of categories, which are described using defined terms instead of
in a free-form manner, into which genes can be assigned. There are
three main super-categories: biological process (BP), molecular
function (MF), and cellular component (CC). Currently, TAIR is
the main curator for GO annotations for Arabidopsis genes, with
some input from other groups. A gene may belong to several
categories and sub-categories at once, which are arranged from
very general to very specific terms (technically, the relationships
between categories and sub-categories are formalized as a directed
acyclic graph).
It is possible to use statistical tests—often a hypergeometric test
or Fisher’s exact test—to assess whether the number of genes
observed associated with a given term (i.e., category) from one’s
list of interest is enriched relative to the number one might expect
to see by chance. Such tests can be used for any classification system
in which objects are classified into categories. Another system of
classification called MapMan Bins was initiated by Bjo ¨rn Usadel and
colleagues at the Max Planck Institute for Molecular Plant Physiology in Germany [51]. This variation on the approach aims to
examine genes whose expression is altered in response to a perturbation in the context of the biological pathways to which they
belong.
3.7.1 AgriGO
AgriGO [52] out of Zhen Su’s laboratory at the Chinese Agricultural University is a user-friendly tool for analyzing whether any
particular GO terms are enriched in a given gene list from Arabidopsis (or for many other agriculturally important species). It provides a nice visualization in the same directed acyclic graph
structure on which the GO system was developed (see Note 17).
1. Go to http://systemsbiology.cau.edu.cn/agriGOv2/ and
select “Analysis Tool” in the tab along the top.
Arabidopsis Bioinformatics
55
sequences to databases already previously described to see if
there is overlap with a known sequence. Otherwise, promoter
deletion experiments in the lab may show necessity of this
sequence for expression, and serve as a starting point for further experimental analyses.
3.7 Functional
Classification
Functional classification of gene lists is one of the basic methods in
bioinformatics for making sense of sometimes rather large gene lists
that arise from gene expression profiling experiments. Typically,
one might look at individual genes in such lists and “see if they fit
biologically,” but one might also like to have an overview of broad
functional categories that change in response to a given stimulus or
due to a specific mutation. One of the very useful large initiatives of
the past decade was the development of a Gene Ontology (GO) for
the “unification of biology” [50]. Basically, this system is based on a
set of categories, which are described using defined terms instead of
in a free-form manner, into which genes can be assigned. There are
three main super-categories: biological process (BP), molecular
function (MF), and cellular component (CC). Currently, TAIR is
the main curator for GO annotations for Arabidopsis genes, with
some input from other groups. A gene may belong to several
categories and sub-categories at once, which are arranged from
very general to very specific terms (technically, the relationships
between categories and sub-categories are formalized as a directed
acyclic graph).
It is possible to use statistical tests—often a hypergeometric test
or Fisher’s exact test—to assess whether the number of genes
observed associated with a given term (i.e., category) from one’s
list of interest is enriched relative to the number one might expect
to see by chance. Such tests can be used for any classification system
in which objects are classified into categories. Another system of
classification called MapMan Bins was initiated by Bjo ¨rn Usadel and
colleagues at the Max Planck Institute for Molecular Plant Physiology in Germany [51]. This variation on the approach aims to
examine genes whose expression is altered in response to a perturbation in the context of the biological pathways to which they
belong.
3.7.1 AgriGO
AgriGO [52] out of Zhen Su’s laboratory at the Chinese Agricultural University is a user-friendly tool for analyzing whether any
particular GO terms are enriched in a given gene list from Arabidopsis (or for many other agriculturally important species). It provides a nice visualization in the same directed acyclic graph
structure on which the GO system was developed (see Note 17).
1. Go to http://systemsbiology.cau.edu.cn/agriGOv2/ and
select “Analysis Tool” in the tab along the top.
Arabidopsis Bioinformatics
55