MetaboAnalyst 5.

0
A Web-based Tool for streamlined
metabolomics data analysis

2022.07.12
 2. Functional Analysis

The Functional Analysis module of MetaboAnalyst has undergone several major updates since it’s introduction in Version 4. First,
it includes a modified Gene Set Enrichment Analysis method, which considers the overall ranks of uploaded peaks and is capable
of detecting more subtle and consistent changes than the original mummichog algorithm (Li et al. 2013). Second, it supports the
inclusion of retention time when performing functional analysis to increase the confidence and robustness of putative compound
annotation. Finally, MetaboAnalyst 5.0 has included an interactive heatmap visualization of a user’s peak intensity table to help
users perform functional interpretation of manually identified patterns of interest.

Other Highlights:

- Users can upload either a peak intensity table (generic or MZMine) or peak list.

- Heatmap based pattern specific functional analysis is available.

- Added support for pathway analysis of 26 organisms including human, mouse, zebrafish, C. elegans, among other species.

- Added ~9, 000 metabolite sets (e.g. Disease-associated sets, chemical classes) to be used for functional interpretation.
 2.0 Knowledge & Background
● Mass spectrometry based untargeted metabolomics traditionally require metabolites to be identified before any biological meaning
can be drawn from the data. Metabolite identification is a challenging and low throughput process, therefore becomes the
bottleneck of the filed. Li et al. report here a novel approach to predict biological activity directly from mass spectrometry data
without a priori identification of metabolites by unifying network analysis and metabolite prediction under the same computational
framework. (version 1)

● The algorithm has been further enhanced to version 2 by considering the retention time information for more accuracy by
introducing empirical compounds. Empirical Compounds are intermediaries between m/z features and compounds. The steps for
how they are formed are as follows:

First, all m/z features are matched to potential compounds considering different adducts. Then, per compound, all matching m/z features are split into
Empirical Compounds based on whether they match within an expected retention time window. The retention time window (in seconds) is calculated
as the maximum retention time * 0.02. This results in the initial Empirical Compounds list.

Next, Empirical Compounds are merged if they have the same m/z, matched form/ion, and retention time. This results in the merged Empirical
Compounds list.

Then, if primary ions are enforced, only Empirical Compounds containing at least 1 primary ion are kept. Primary ions considered are 'M+H[1+]',
'M+Na[1+]', 'M-H2O+H[1+]', 'M-H[-]', 'M-2H[2-]', 'M-H2O-H[-]', 'M+H [1+]', 'M+Na [1+]', 'M-H2O+H [1+]', 'M-H [1-]', 'M-2H [2-]', and 'M-H2O-H [1-]'. This
results in the final Empirical Compounds list.

Finally, pathway libraries are converted from "Compound" space to "Empirical Compound" space. This is done by converting all compounds in each
pathway to all Empirical Compound matches. Then the mummichog/GSEA algorithms work as before to calculate pathway enrichment.
2.1 Start Functional Analysis

Click here to
start
2.2 Starting from a list

From peak list to pathways

 2.2.1 Peak Uploading – peak list
TIP1 : Multiple examples
are provided here. Please
try to download one and
follow the style of the
example rigorously. Please
make sure that the list
header is consistently same
as the example.

2. Set parameters (mass error

and ion mode) based your
instrument

1. Switch the different

uploading data type from
here (peak list or table)

3. Click submit to
continue
2.2.2 Data Integrity Check

1. Check Data
Integrity Result to
make sure correct

2. Click Proceed to
continue
 2.2.3 Set Parameters

Customize the Metabolites

Set corresponding
parameters/Library

Customize the Adducts

TIP1 : You can manually TIP2 : Heatmap analysis

customize the abundant only works if users upload a
substances as ‘Currency peak table If you want to do
Metabolites’ and adducts the heatmap based
for not considering at the analysis, please see 6.3.
pathway analysis.
2.2.4 Pathway analysis results
1. The compounds/empirical
compounds hits in this pathway

2. Click buttons at the bottom of this

page to download results or go the
network exploration page
2.3 Starting from a table
 2.3.1 Peak uploading – peak table
TIP1 : Currently, 2 types of
peaks are supported
(MetaboAnalyst generic
and MZmine style). Users
could coerce you table
manually to make it
applicable here.

2. Set parameters (mass error and

ion mode) based your instrument

1. Switch to uploading
peak intensity table tab

3. Click submit to
continue
 2.3.2 Peak uploading – Preprocessing

1. Perform Data 2. Perform Data 3. Perform Data

Integrity Check Filtering Normalization
2.3.3 Set parameters
TIPs : Most parameters are
same as the ones used for
processing the peak list, as
described in 6.2.3. The only
difference is that peak table
allow the heatmap based
pattern specific functional
analysis.
Set corresponding
parameters/Library
2.3.4 Heatmap based pattern specific analysis
TIPs : The scatter plot and
its corresponding functions
for peak table uploading is
totally same the one of
peak list uploading. Please
refer to 6.2.4.

1. Select Heatmaps
radio to start !
2.3.5 Heatmap based pattern specific analysis - result
This section maybe too complicated to easily understand/follow for beginners, why not watch a video first?

Read these tips and

click Ok to start
 2.3.6 Heatmap interface introduction
Controller panel to
adjust parameters
for clustering.

Overview of
the peak
across the
spectrum.
Default is Pattern based
clustered Enrichment
based on p analysis panel.
value.

Focus view of
the specific peak
pattern from the
whole spectrum. Dynamic Display panel,
Default is the top used to show the
Sample
50 peaks. peak/sample information
Names dynamically.
View panel.
 2.3.7 Heatmap peak clustering

3. The selected
1. Select a clustering specific spectral
method from Cluster peaks peaks’ pattern
menu, e.g. Ward’s method appears in the
focus view panel.

TIP1 : The clustering will be 2. Left click your mouse, don’t

performed once you select release and move your cursor
the corresponding methods. over a certain area across the
Clustered peak heatmap will spectrum until finish. Then TIP2 : In present example,
be updated immediately. we selected a consistently
release the left click.
down-regulated pattern in
‘green’ group, and
consistently upregulated
in ‘red’ group.
2.3.7 Peak patterns’ stitch -1

2. Hold the left

click of your
1. If you want to stitch mouse and move
the different peak over the whole
patterns. Click the area as the
Builder Button and the selected part.
stitch tools appears. The selected part
will appear at the
bottom panel.

Continue at the next page

 2.3.7 Peak patterns’ stitch -2

4. New pattern
appears immediately
at the top focus view
panel. Hold your left
click mouse and move
over from this focus
view to confirm the
area.

6. Click To
Focus view.
3. Left click and move
your mouse over a
new pattern from the 5. Newly selected area will be
overview panel. stitched with the previously 7. The stitched peaks’
selected pattern at the pattern will be presented
bottom focus view. at the Focus view.
2.3.8 Enrichment Analysis -1
TIP1 : Operation Mode could
show the hits in different way.
Annotate will annotate directly
1. Select on the heatmap, while the
database and ‘Extract’ will extract the ions
submit to do hits and hide other non-hits.
the enrichment
analysis

Continue at the next page

2.3.8 Enrichment Analysis -2

2. Select the pathways

you are interested in,
then the corresponding
hits will appear at the
right side of the Focus
view panel.
 2.3.9 Result Download

1. Set the resolution of

the image in Focus
view 2. Download the
images from Download
menu and click to
download it!
 2.3.9 Result Download

Click the “Generate

Report” to download a
pdf report summarizing
your analysis.
 Thanks

If you have any questions please read through the FAQs or contact us at
Zhiqiang.pang[at]xialab.ca or Jeff.xia[at]xialab.ca