enabling the analysis of more than 10^4 samples
per year.
Recently, the accelerated increase, in many
species, in plant resources for functional genom-
ics, such as wild accessions, insertional knockout/
overexpresser lines and recombinant inbred lines
(RILs), has made high-throughput plant meta-
bolomics ever more necessary. As a case study, a
widely targeted analysis of amino acids and amino
acid-derived secondary metabolites was used to
screen approximately 3,000 transposon inser-
tional knockout lines and wild accessions of
Arabidopsis for over- and under-accumulation
mutants of specific metabolites (Nakabayashi
et al.2013a). As a result, one line was shown to
accumulate branched-chain amino acids (valine,
leucine and isoleucine) in up to 100 times the
amount of the wild type of the background eco-
type. Interestingly, this mutant phenotype may be
specific to that accession; no mutant that overac-
cumulates branched-chain amino acids has been
found in T-DNA insertion lines provided by the
Salk Institute for Biological Studies. Character-
ization of such natural variation is important for
elucidating plant metabolic regulation. Moreover,
quantitative locus (QTL) mapping using RILs is
expected to facilitate the identification of novel
genes without reference to sequence homology.
Widely targeted metabolic profiling will assist
metabolic QTL (mQTL) studies, which are
effective in assigning enzymes and regulatory
genes to the network of known metabolic reac-
tions (Lisec et al. 2008 ; Brotman et al. 2011 ).
16.5 Integrated Metabolomics
with Recombinant Inbred
Lines ofL. japonicus
L. japonicus is used for functional genomic
approaches in studies of leguminous metabolism
because its genomic infrastructure has facilitated
the identification and characterization of bio-
synthetic genes (Shimada et al. 2005 , 2007 ;
Forslund et al. 2004 ; Morant et al. 2008 ; Saito
et al. 2012 ). Accumulating DNA markers and
SNPs of the model legume is making mQTL
analysis feasible, in which the metabolic profile
is used as a quantitative phenotype for QTL. This
analysis is expected to address the gaps between
genomics and metabolomics. An integrated
metabolomics platform consisting of a combi-
nation of widely targeted and untargeted analyses
was established with the aim of discovering the
comprehensive mQTL set ofL. japonicus. The
putative metabolite structures associated with
significant mQTLs were assigned by a MS/MS
database search for phytochemicals (Sawada
et al. 2012 ; Sawada and Hirai 2013 ).
As described above, two UPLC-MS/MS-based
platforms for plant metabolomics are available:
untargeted analysis using UPLC-QTOF-MS and
widely targeted analysis using UPLC-TQ-MS.
UPLC-QTOF-MS for untargeted analysis
acquires the data of molecular mass and MS
spectra for virtually all LC-separated peaks, but
they are redundant for each peak. Untargeted
analysis thus yields a vast amount of data (a few
gigabytes per analysis) and can be applied only
with difficulty to large-scale analyses dealing with
hundreds to thousands of samples. In contrast, in
SRM using UPLC-TQ-MS for widely targeted
analysis, no MS spectrum is recorded, and the
transitionand other parameters for each compound
are optimized in advance. Accordingly, SRM
using UPLC-TQ-MS yields data sets of practical
size (a few megabytes per analysis) and provides a
high-throughput metabolomics platform.
Given that high-sensitivity detection by TQ-
MS requires optimization of analytical conditions
with authentic standard compounds, the next
challenge of widely targeted analysis is to detect
unidentified metabolites. For this purpose, the
SRM conditions for unknown compounds may be
best obtained by conversion of the data acquired in
an untargeted analysis with UPLC-QTOF-MS,
such as MS2Ts. In the new integrated method,
data conversion is implemented using a newly
developed program, and unidentified compounds
corresponding to MS2Ts are analysed using SRM
conditions that are converted from MS2Ts, named
integrated SRM (iSRM). In the case study inL.
japonicusseeds (B-129 and MG-20), a total of
80554 MS2Ts were collected in positive and
negative ion modes. The SRM conditions derived
from the MS2Ts were tested using the MS peak16 Metabolomics 177