Metabolomics Identification Report Standards Version 2

A Metabolomics unknown feature identification report industry standards from BioNovoGene corporation.

#2019.08.16# at Suzhou, China

Table of Content

• 1. Definition and background
• 1. Report Table Format
     • 2.1. section1: basic info of ion feature
     • 2.2. section2: annotation information
     • 2.2.1. part1. the basic meta information
     • 2.2.2. part2. the external database cross reference id
     • 2.2.3. part3. the chemical structure information of target * 2.3. section3: annotation score
• 1. Alignment Visual
• 1. Supplementary

1. Definition and background

There is a general consensus that supports the need for standardized reporting of metadata or information describing large-scale metabolomics data sets. Reporting of standard metadata provides a biological and empirical context for the data, enables the reinterrogation and comparison of data by others, which is also could let us interpret the result in a more clearly way.

This article is mainly address at the unknown metabolite identification in LC-MS experiment, and proposes the reporting standards related to the chemical analysis aspects of metabolomics experiments its metabolite identification.

Some terms in this article that address to:

• feature, the term feature in this article is refer to a parent ion in LC-MS experiment result raw data. Where a parent ion feature is a peak in chromatography data, which is consist of mass to charge ratio in ms1 level and its retention time (with a range of lower bound and upper bound) in chromatography experiment result.
• annotation, the term annotation in this article is refer to the multidimensional information about the metabolite that assigned to a unknown feature, which such multidimensional information consist with the metabolite its cross reference id in different database, common name, basic chemical data like mass and formula composition and its molecule structure information, etc.
• alignment, the term alignment means a kind of operation that use to compare the similarity of the mass spectrum data between user sample and the reference standard library. Such similarity comparison result is the most important evidence that use for unknown feature its identification.
• score, the term score is a kind of numeric value that produced by the alignment comparison calculation. Literally, the higher score the alignment it produce, the better the result it is.

Our metabolite identification report consist with two parts of data which present to our user:

1. Report excel table that contains the raw sample information and the meta annotation information of the metabolite.
2. Data visual plot for the mass spectrum alignment details.

2. Report Table Format

2.1. section1: basic info of ion feature

2.2. section2: annotation information

In this section, includes three parts of annotation information to describ the resulted metabolite.

2.2.1. part1. the basic meta information

2.2.2. part2. the external database cross reference id

2.2.3. part3. the chemical structure information of target

2.3. section3: annotation score

This section of data consist with two parts of information. First part of the information is the alignment information for the metabolite identification, includes: the reference standard source trace in the reference library, and multiple dimensions of the scores of current identification.

And then, is the final identification confidence result for current alignment.

The result of identify.level have literal values for the representation of the identification its confidence level(defined by BioNovoGene):

• confirm: the unknown feature is confirm as the assigned metabolite both in ms1 level and ms2 level
• MSMSconfirmed: the unknown feature is confirm as the assigned metabolite in ms2 level, and ms1 m/z ratio, but not match in rt relative adjustment result value.
• MSMScheck: the unknown feature is probably confirm as the assigned metabolite in ms2 level and ms1 m/z ratio, but have low score in ms2 alignment and rt also not match.
• ms2hit: the unknown feature is probably confirm as the assigned metabolite, but have low score in ms2 alignment and rt match, we only sure that the alignment have the relative high number of shared fragments between the user sample and the reference library.
• ms1hit: the unknown feature didn’t have the ms2 level hits in the reference spectrum database, but has the alignment reuslt both in dimensions m/z and rt within the given tolerance value.

The result of MSI_levels have literal values for the representation of the identification its confidence level:

+

The metabolomics identification result also can be divided into

• Actually identified as
• **Chemical Character ** + +

References

The work of this metabolomics identification report standards from the BioNovoGene R&D labs is based on the research works of:

1. SSM algorithm is original defined in this literature article: “MetDIA: Targeted Metabolite Extraction of Multiplexed MS/MS Spectra Generated by Data-Independent Acquisition.” DOI: 10.1021/acs.analchem.6b02122
2. MetaDNA algorithm is original defined in the literature article: “Metabolic reaction network-based recursive metabolite annotation for untargeted metabolomics.” DOI: 10.1038/s41467-019-09550-x
3. pvalue for the metabolites’ identification is tested from the BHR score vector, and then test by the in-house hyper-geometric algorithm which is developed by BioNovoGene. The BHR vector which is original defined in this literature article: “KAAS: an automatic genome annotation and pathway reconstruction server.” DOI: 10.1093/nar/gkm321.
4. FDR.ratio or FDR.flag is the xyMeta algorithm result for . which is : “XY-Meta: A high-efficiency search engine for large scale metabolome annotation with accurate FDR estimation.” DOI: 10.1021/acs.analchem.9b03355
5. rt.adjust is the score result of the scalable chromatography mapping result, which is generated by a machine learning method from the literature article: . The method is laterly improved in BioNovoGene R&D labs.
6. MSI_levels “Proposed minimum reporting standards for chemical analysis”. DOI: 10.1007/s11306-007-0082-2
7. identify.level confidence levels is defined by BioNovoGene, this identify.level concepts is developed from this literature article: “Identifying Small Molecules via High Resolution Mass Spectrometry: Communicating Confidence”. DOI: 10.1021/es5002105

3. Alignment Visual

Due to the reason of mass spectrum alignment is the most important part in unknown metabolite identification, so it is important to make a clear presentation of the alignment result to user.

The mass to mass spectrum alignment reuslt can be visual by bi-direction barplot, where the top part of the plot which in blue is the ms2 mass spectrum in user sample; and the bottom half of the plot which in brown color is the ms2 mass spectrum of the reference in standards library.

Only the molecule fragment its intensity value greater than 30%, then its mz ratio will be display on this barplot.

4. Supplementary