Interface reference

1. Click the upload button and select the file you want to convert.
2. Choose whether to use the file directly as input or convert it to BiomiX format. If you select "Yes" when asked "Do you want to modify the file?", a second dialog will appear asking for the file separator, header, ID column, and decimal separator.
3. The Preview tab shows the first 10 columns and rows of the matrix. Use the editing tab to remove specific columns or rows, or to transpose the matrix if needed.
4. Once satisfied, confirm to use the converted matrix as input.

Selecting Preview-QC in an omics input slot opens the BiomiX toolkit, a Shiny interface with four sections:

Section	Description
Upload Data	Upload data for imputation. Available methods: replacement by 0, variable mean or median, lasso, NIPALS, and random forest. The filter button removes samples and variables exceeding the specified threshold of 0 or missing values.
Boxplot by variables	Shows the percentage of 0 or missing values per variable.
Boxplot by samples	Shows the percentage of 0 or missing values per sample.
Summary table	Visualizes the modifications applied during imputation.
Download	Download the imputed and transformed matrix.

General parameters

Parameter	Description
Condition	The group to analyse.
Control	The reference group, compared with the condition group.
Output	Directory where analysis results will be saved.

Omics input grid

Parameter	Description
Input	Input slot number.
Preview-QC	Opens the BiomiX toolkit Shiny app for data visualization, transformation, and outlier removal.
Single omics Analysis	Check to run single-omics analysis on this layer using the BiomiX pipelines.
Data type	Select the omics type of the uploaded data.
Integration	Check to include this omics layer in the integration analysis.
Label	Omics name, used to name the output folder and to filter samples based on sample names.
Selection	If selected, filters samples based on sample name using a regex derived from the label name.
Data upload	Upload the omics matrix. Accepted formats: .tsv, .xls, .xlsx. The format converter is also accessible here.

Integration grid

Parameter	Description
Integration	Enable multi-omics integration with the uploaded omics data.
Method	Select the integration method (MOFA, SNF, NEMO, DIABLO).
N° Factors	Number of MOFA factors to compute. Set to 0 for automatic selection.
Factor to explore	Which factor to visualize graphically (contributors, heatmap clustering, etc.).
Omics overlay	Minimum number of omics layers a sample must have to be included in the integration.
Open advance option	Opens the advanced options interface.
Open BiomiX chatbot	Opens the BiomiX chatbot for guided assistance.
Start Analysis	Launches the analysis with the current configuration.

Parameter	Description
Log2FC threshold	Log2FC threshold value for calling significant results.
P.adj threshold	Adjusted p-value threshold for calling significant results.
Gene Panel	Upload a gene panel file for subgrouping analysis. Genes must be in Gene Symbol format, tab-delimited.
Array type	Type of methylation array (450K or EPIC).
N° genes with score > 1 or 2	Defines the positive/negative subgrouping threshold: number of panel genes that must have a Z-score above one or two standard deviations of the control.
Remove positive controls	If checked, control samples positive for the gene panel are excluded from downstream analysis.
N° top DE genes in heatmap	Number of top differentially expressed genes (or metabolites) to show in heatmaps.
Clustering distance	Distance metric used for clustering in heatmaps (subgrouping and single-omics).
Clustering method	Clustering method used in heatmaps.
CPU threads	Number of CPU threads to use in parallel processing.
N° MOFA input features	Number of top features used as MOFA input. A similar number per omics layer is recommended.

Parameter	Description
Metabolite annotation	Identifier type used in the uploaded annotated metabolomics matrices (HMDB, KEGG, or compound name).
Ion mode	Ionization mode used during data acquisition.
M/Z Tolerance ppm	PPM tolerance for MS1 annotation.
Adduct positive mode	Type of positive adducts generated during acquisition.
Adduct negative mode	Type of negative adducts generated during acquisition.
Adduct neutral mode	If checked, treats data as neutral mode.
MS1 files upload	Upload the MS1 annotation file with the corresponding input slot number.
Databases MS1	Databases queried for MS1 annotation via CEU Mass Mediator (HMDB, LipidMaps, Metlin, KEGG).
mz match MS1/2	Mass/charge ratio tolerance for matching MS1 annotations against MS2 fragmentation spectra.
RT match MS1/2	Retention time tolerance for matching MS1 annotations against MS2 fragmentation spectra.
Column	Type of chromatography column used in liquid chromatography.
Databases MS2	Databases queried for MS2 annotation via TidyMass (HMDB, MassBank, MoNA).
MS2 directory	Directory containing fragmentation spectra files (.mzML or .mgf).

Parameter	Description
Column name	The metadata column to use for sample filtering.
Data type	Type of data in the selected column (numerical or categorical).
Threshold / Factor	Threshold for numerical filtering (e.g., >= 90 or == 0.56) or category for categorical filtering (e.g., == male or == treated).

Parameter	Description
Max iteration	Maximum number of MOFA iterations during model training.
Convergence mode	Training speed. Slower modes produce more accurate models.
Threshold contribution weight	Threshold to isolate top MOFA factor contributors. Among the top 5% contributors (pre-filtered by BiomiX), defines a contribution threshold (default: 0.5).
Type of research (Bibliography)	Document type for PubMed search: abstract/title or full text word.
N° articles (Bibliography)	Number of PubMed articles consulted per factor. High values may trigger PubMed rate limits — reduce if the search is interrupted.
N° top contributors (Bibliography)	Number of top contributors per omics searched in PubMed abstracts.
N° keywords extracted (Bibliography)	Number of keywords extracted per article by text-mining (litsearchr). These are the most cited word combinations in the retrieved abstracts.
P.adj threshold (Pathway mining)	Adjusted p-value threshold to consider a biological pathway significant.
Pathways shown (Pathway mining)	Number of biological pathways visualized in the PDF reports.
Numerical (Clinical)	Correlate numerical clinical data with significant MOFA factors (Pearson correlation).
Binary (Clinical)	Correlate binary clinical data with significant MOFA factors (Wilcoxon test).
Save advanced options	Click the blue button to save the current advanced parameters before starting the analysis.