raw data collected from different experiments, the
quantile method with Robust Multichip Average
(RMA) was used (Irizarry 2003 ). The raw data
(.CELfiles) were also imported into dCHIP soft-
ware (Li and Wong 2001 ) to assign presence/
absence calls for each probe set using the soft-
ware’s default settings.
17.4.1 Search Criteria
To make LjGEA user-friendly, many different
options are given so that even partial information
about a given gene is enough to serve as a search
criteria and the corresponding transcript can be
retrieved (Fig.17.1). The three main types of
queries that can be used tofind expression data
for a specific gene are as follows:
A. The probe set ID can be used. It is the
quickest and the most accurate way of
retrieving expression data, given the probe
set is specific to the gene. The limitation, on
the other hand is that the user has to know the
exact probe set ID, which is often not pos-
sible with just a sequence in hand.
B. The gene ID or gene annotation: If known, a
gene ID (according to theL. japonicusGen-
ome Database or LjGDB) can be used as a
query. Alternatively, a gene description (e.g.,
ABC transporter), gene annotation or gene
ontology (GO) (e.g., GO: 0004757 for
sepiapterin reductase), and KEGG gene bin-
codes (e.g., 1.1.1.7, which stands for plastidic
aldolase) can also be used. The query must be
given in specific formats, which are shown as
examples on the server.
C. BLAST search: This can be used when only
the sequence of the gene or a part of the
sequence of the transcript or cDNA sequence
is known. This can also be useful for an
unknown gene or a gene of which the
ortholog from other species (e.g., Medicago)Fig. 17.1 The Lotus GEA Web server: On theleft panel,
clickable links open search pages. The search criteria can
be probe set IDs, gene IDs, gene description, GO terms, or
KEGG terms. All the search pages contain examples of
the correct format for inputting queries192 J. Verdier et al.