SE51:/S01/M01/D01
Sample Set Information
| ID | SE51 |
|---|---|
| Title | RIKEN tandem mass spectral database (ReSpect) for phytochemicals: A plant-specific MS/MS-based data resource and database |
| Description | The fragment pattern analysis of tandem mass spectrometry (MS/MS) has long been used for the structural characterization of metabolites. The construction of a plant-specific MS/MS data resource and database will enable complex phytochemical structures to be narrowed down to candidate structures. Therefore, a web-based database of MS/MS data pertaining to phytochemicals was developed and named ReSpect (RIKEN tandem mass spectral database). Of the 3595 metabolites in ReSpect, 76% were derived from 163 literature reports, whereas the rest was obtained from authentic standards. As a main web application of ReSpect, a fragment search was established based on only the m/z values of query data and records. The confidence levels of the annotations were managed using the MS/MS fragmentation association rule, which is an algorithm for discovering common fragmentations in MS/MS data. Using this data resource and database, a case study was conducted for the annotation of untargeted MS/MS data that were selected after quantitative trait locus analysis of the accessions (Gifu and Miyakojima) of a model legume Lotus japonicus. In the case study, unknown metabolites were successfully narrowed down to putative structures in the website. |
| Authors | Yuji Sawada, Ryo Nakabayashi, Yutaka Yamada, Makoto Suzuki, Muneo Sato, Akane Sakata, Kenji Akiyama, Tetsuya Sakurai, Fumio Matsuda, Toshio Aoki, Masami Yokota Hirai, Kazuki Saito |
| Reference | Sawada Y et al. (2012) Phytochemistry 82: 38-45 |
| Comment |
The raw data files are available at DROP Met web site in PRIMe database of RIKEN.
Sample Information
| ID | S01 |
|---|---|
| Title | L. japonicus accessions |
| Organism - Scientific Name | Lotus japonicus |
| Organism - ID | NCBI taxonomy:34305 |
| Compound - ID | |
| Compound - Source | |
| Preparation | The L. japonicus accessions (MG-20 and B-129) and mapping population of 129 recombinant inbred lines (RILs) have been described in detail elsewhere (Gondo et al., 2007; Hayashi et al., 2001; Kawaguchi et al., 2001; Klein and Grusak, 2009). |
| Sample Preparation Details ID | |
| Comment |
Analytical Method Information
| ID | M01 |
|---|---|
| Title | Metabolic profiling |
| Method Details ID | MS01 |
| Sample Amount | N.A. |
| Comment | Not written in the paper. |
Analytical Method Details Information
| ID | MS01 |
|---|---|
| Title | Metabolic profiling and mQTL analysis |
| Instrument | UPLC-QTOF Premier (Waters) and UPLC-TQS (Waters) |
| Instrument Type | UPLC-QTOF-MS |
| Ionization | ESI |
| Ion Mode | Positive and Negative |
| Description | <Metabolic profiling>
A seed was disrupted using a multi beads shocker (Shake Master NEO, BMS, Tokyo, Japan), and the seed powder was extracted with 1 mL of extraction buffer (0.1% HCO2H, MeOH–H2O (4:1 v/v), and 33.6 nM lidocaine and 840 nM camphor sulfonic acid as internal standards of positive mode and negative mode, respectively) using a multi beads shocker. After centrifugation (4 °C, 10,000 rpm, 5 min), the sample tubes were subjected to sample preparation (buffer transfer, 250 μL of seed extract; dried up, resolution, 250 μL of LC-MS grade H2O; filtration, 384 well formatted filter (0.45 μm PVDF, Whatman, NJ, USA) with a liquid handling system (Microlab Star Plus, Hamilton, Ontario, Canada) (Sawada et al., 2009a). The metabolites were analyzed using a liquid chromatography quadrupole time of flight mass spectrometer and tandem quadruple mass spectrometer system (UPLC-QTOF Premier and UPLC-TQS, Waters Co., MA, USA) (Matsuda et al., 2009; Sawada et al., 2009a). The metabolites were analyzed using a liquid chromatography quadrupole time of flight mass spectrometer and tandem quadruple mass spectrometer system (UPLC-QTOF Premier and UPLC-TQS, Waters Co., MA, USA) (Matsuda et al., 2009; Sawada et al., 2009a). After cutting off the low intensity data of MS2Ts, 702 MS2Ts were used for a selected reaction monitoring (SRM) assay of UPLC-TQS. To determine the optimal collision energy for MS2Ts, collision induced dissociation fragmentation analyses at six energy steps (10–60 eV) were conducted using the same seed extracts. As a result, 4212 SRMs were used for analysis. The SRM condition that gave the highest signal intensity among the six energy steps was defined as the optimal SRM. The optimal SRM conditions were further selected based on the following criterion: the relative standard deviation of UPLC-TQS peak area values must be less than 10% (analytical replicates = 3). Finally, 342 SRM conditions for MS2Ts were successfully assigned to the seed extract of L. japonicus. Based on the natural variations in the metabolite accumulation patterns among the accessions (MG-20 and B-129), 88 SRM conditions were used for mQTL analysis.
Metabolic profiling data of 129 RILs were processed as follows: the peak area values of 88 SRMs were divided by the peak area values of the internal standards, which were lidocaine and camphor sulfonic acid for the positive and negative ion modes, respectively. The divided values were then converted into z-scores of binary logarithms after missing values were replaced with 0.1. The genotype data of each RIL were obtained from KDRI (http://www.kazusa.or.jp/lotus/). Using processed metabolic profiling data and gene marker data of RILs, mQTL analysis was carried out using R/qtl (Broman et al., 2003) (Supplementary Data). |
| Comment_of_details |
Data Analysis Information
| ID | D01 |
|---|---|
| Title | ReSpect |
| Data Analysis Details ID | DS01 |
| Recommended decimal places of m/z | Default |
| Comment |
Data Analysis Details Information
| ID | DS01 |
|---|---|
| Title | ReSpect |
| Description | <MS/MS data in ReSpect>
The literature used for the MS/MS data were obtained from PubMed and Google Scholar, and the MS/MS data were manually digitalized. All literature used are listed in the information section in ReSpect (http://spectra.psc.riken.jp/). For studies in which the relative intensity value of the MS fragment spectra was not described digitally, the heights of the MS peaks in the figures were manually measured and calculated. For literature data that have only m/z values without intensity data, the intensities of all ions to 100% were adjusted. In addition, the MS/MS data of the authentic standard compounds were acquired using Q-TOF/MS and QqQ/MS (Matsuda et al., 2010, 2009; Sawada et al., 2009a,b,c). All records used for ReSpect were categorized into compound classes according to the book (Cane, 1999; Kelly, 1999; Pinto, 1999; Sankawa,1999). The records were formatted using MassBank Record (Horai et al., 2010), which can accept MS/MS data as well as its metadata (summary, chemical, analytical, spectral and outer database links). As of September 2011, ReSpect contained 8649 records corresponding to 3595 metabolites. Specifically, the database contains 163 selected literature reports (3341 records corresponding to 2741 metabolites), Q-TOF/MS data (1050 records corresponding to 575 standard compounds) (Matsuda et al., 2010, 2009) and QqQ/MS data (4258 records corresponding to 861 standard compounds) (Sawada et al., 2009a,b,c). The ratios of the metabolite ontology classes are biased toward the available data and referenced classification of metabolites. In the data section of ReSpect, the example analyses (statistics and evaluation) were automatically executed based on the updated records.
In this study, association rules and index values (transaction, support, confidence, lift) of MS/MS data in the ReSpect records were enumerated using a module of Perl (Data Mining Association-Rules Version 0.10). To mimic the MS/MS data structure, the maximum value of the product ion and other product ions are defined as the antecedent and consequent, respectively, an antecedent is defined as a single ion, and the difference in m/z values between the antecedent and consequent is greater than 10. Moreover, a rare rule (only once in the transaction) was also excluded as invalid rules. On the basis of these definitions, the following invalid rules were excluded from the association rules: multiple antecedents in a rule, antecedent < consequent, and the estimated ion of the consequent corresponding to the isotopic ion of the antecedent. The index values of MS/MS fragmentation association rules with an antecedent (X), consequent (Y) and transaction (Z) are calculated as follows: support (X) = count (X)/count (Z), support (Y) = count (Y)/count (Z), support (X)Y) = count (X [ Y)/count (Z), confidence (X)Y) = count (X [ Y)/count (X), lift (X)Y) = confidence (X)Y)/support (Y). The MS/MS fragmentation association rules can be downloaded from the data section of ReSpect.
The web applications of ReSpect were implemented using HTML5.0, JavaScript, Perl, and MySQL and were hosted on CentOS. ReSpect has been successfully tested on Mozilla Firefox 3.6+, Safari 5.0+, Google Chrome 5.0+, Opera 10.0+, and Microsoft Internet Explorer 8.0+. The user can define the threshold for query and data resources with the following search options: polarity (positive, negative), spectra filters (similarity score, relative intensity, precursor ion and m/z tolerance), data type selection (ignores all single fragments and one major fragment spectrum) and data set selection (literature, Q-TOF/MS and QqQ/MS). |
| Comment_of_details |
