Usage

The Dashboard is the main page of FERMO GUI. It allows the perform data inspection and analysis and to download generated files for further use.

Examples

Three examples are available that illustrate the prioritization capabilities of the FERMO dashboard.

Euphorbia dendroides

This example re-analyzes the Euphorbia dendroides dataset previously featured in the Bioactivity-Based Molecular Networking publication. This dataset integrates metabolomics data (MSV000080502) with quantitative phenotype measurements of antiviral activity against the chikungunya virus.

This example demonstrates that FERMO’s phenotype prioritization approach yields results consistent with those reported in the original study. Applying the phenotype filter prioritizes the following molecular features: m/z 591.324 (RT 27.05 min), m/z 589.311 (RT 25.4 min), and m/z 563.296 (RT 21.6 min), highlighting FERMO’s ability to reproduce biologically relevant findings despite using a different prioritization algorithm.

Streptomyces sp. MBT27

This example revisits a study, where cultivation on various carbon sources led to differential antibacterial activity against Staphylococcus aureus. The dataset integrates metabolomics data (MSV000097564), quantitative bioactivity data, growth condition metadata, and genomic information (antiSMASH results).

This example showcases FERMO’s strength in integrating diverse data types to identify bioactive compounds. In this case, FERMO correctly prioritizes actinomycin D as the active compound, further linking it to its likely biosynthetic gene cluster.

Planomonospora

This example analyzes a subset of 10 Planomonospora strains, originally described in this study. This dataset combines metabolomics data (Zenodo) with qualitative bioactivity data (growth inhibition against Staphylococcus aureus) and information about the phylogroup to which each strain belongs.

This case demonstrates FERMO’s capability to integrate bioactivity and phylogenetic context. Applying the phenotype filter reveals that only strains from the “S” phylogroup exhibit antibacterial activity. Feature inspection suggests that the activity is likely due to siomycin A and its congeners, well-known antibacterial thiopeptides.

Dashboard Elements

Nota bene: we will illustrate the dashboard functionality by drawing from the Planomonospora example dataset

Main chromatogram

The Main chromatogram view is the core of the data visualization. It allows to access information on molecular features, their annotations, and relatedness. Clicking a peak triggers the sequence similarity visualization as chromatogram (showing releated features in the active sample) and as network (showing related features in the whole dataset.)

Nota bene: the peaks (extracted ion chromatograms) are actually pseudo-chromatograms, reconstructed from the peaktable data. They are only an abstraction of the original data and may not always represent the original peak shape appropriately.

Example

In the example, we inspect the molecular feature with the ID 15 by clicking it. This feature has multiple related features present in the sample and in the dataset. When we inspect the Molecular feature info we see that it has been annotated as chymostatin-like molecule.

General information

The General information displays a summary of the analysis run.

Example

In the example, we see that 143 molecular features from 11 samples were processed. 74 features were removed due to the filter settings. We can also see the run date and the version of fermo_core that was used.

Sample overview

The Sample overview gives an overview of the samples in the analysis. It also shows the Retained features, that is, which features have been selected by the filter settings. Further, it displays sample scores and group affiliation of the samples. A click on a sample name changes the visualization in the Main Chromatogram.

Example

In the example, we click through some of the samples. Since we have not set any filters, the numbers in Total features and Retained Features are identical. We then look for the sample with the highest diversity, specificity, and mean novelty score.

Molecular feature info

The Molecular feature info gives more information on the selected molecular features. This includes general information, annotation information, abundance across groups and samples.

Example

In the example, we inspect the feature ID 83. We see that it has been annotated as a siomycin-like compound. It also has an annotation to be phenotype-associated. Further, it has been annotated by the Adduct, Fragment, and Neutral loss modules.

Spectral similarity networking

The Spectral similarity networking displays the spectral similarity network (also known as molecular networks) the selected feature is associated with. FERMO can calculate these networks with different algorithms, which result in different network topologies.

Example

In the example, we inspect a network that shows very different topologies dependent on the algorithm used. We inspect the nodes and edges, and use the network view to quickly cycle through the features.

Filters

The Fiters are a useful element of the data analysis: they allow to quickly focus attention on the most relevant features for the respective research questions. Filters can be stacked to allow for specific research questions. The filter settings also update the Retained features number in the Sample overview.

Nota bene: we are currently working on further improving filter settings.

Novelty score filter

Allows to filter for the Novelty score.

Hide blanks

Allows to hide blank-annotated features (see here).

Hide unannotated features

Allows to hide all features that do not have any annotations. This also triggers the availability of additional filters.

Match score filter

Allows to filter for (spectral) library matches in a certain range.

Phenotype score filter

Allows to filter for a Phenotype score in a certain range (only showing phenotype/bioactivity-associated features).

Find feature

Allows to search for a specific feature ID.

Precursor m/z filter

Allows to search for a precursor m/z range

Sample filter

Allows to select features that have been observed in a minimum and/or maximum number of samples.

Fold-change filter

Allows to select features that are n-times more abundant across one group versus another group in a certain category.

Show only selected group feature

Allows to show only features that are in a selected group

Exclude feature with specific network groups

Allows to exclude features that are associated to a certain group from the selection. This also extends to all features that are in the same spectral similarity network. This can be useful to subtract blank-associated features.

Example

In this example, we go through all the filters and observe how they impact the selection.

Downloads and general settings

This field allows to download the individual files for storage or offline use. For more information about the output files, see here.

Display

In FERMO v.1.0.0, the dashboard has been completely reworked and now allows for dynamic construction of the user interface.