Annotate Your Model

Data and models in Systems Biology can come in various formats. MIRIAM-compliant resource annotations are a practical, standardised way to connect the data or model elements with a biological meaning (e.g., metabolites, proteins, genes). This is done by linking them to entries in public biological databases, ontologies, or other web resources listed in the MIRIAM resources. In the SBML format for models, this is already widely used.

The online service "Annotate Your Model" helps you to add such annotations to biological data and to models in non-SBML formats. This works as follows:

1. Upload your data or model file.
   More information on supported file formats

   You can upload data or model files in different formats. The elements to be annotated (data rows or model variables) have to be specified.

   
   

   1. Data tables. You can upload omics datasets as Excel .xls, .tsv, or .csv files. Please note that the extension of the file you upload is critical! If the extension is .csv or .tsc, commas or tabs will be used as a seperator. After uploading a file, you will be asked to select the table column listing the elements to be annotated (e.g., compound or gene names).

   
   

   2. Model files. You can annotate model files in the following formats:

   • Matlab

     % model variables
     % var1,var2,var3
     % end variables

   • R

     # model variables
     # var1,var2,var3
     # end variables

   • Java

     // model variables
     // var1,var2,var3
     // end variables

   • C/C++

     // model variables
     // var1,var2,var3
     // end variables

   • Python

     # model variables
     # var1,var2,var3
     # end variables

   Model elements have to be specified in inline comments within the model code (see the shaded box, select file format with mouse). Model files are not uploaded directly, but in the form of a .zip archive, which can contain several model files. If it also contains a MIRIAM.csv file created earlier with "Annotate Your Model", the annotations in this file will be read and can be edited in the following.

   
   

   
   
2. You will be shown the list of elements to be annotated (i.e., the variables of a model or the rows of a data table) and can add more elements by using the box "Add Elements".
3. After clicking an element in the list, you can add or edit its annnotations.
4. Once you have annotated your model, you can
   • export the annotations in a CSV file.
     More information on annotation format

     MIRIAM-compliant external resource annotations are a practical, standardised way to connect biological data or model elements with a biological meaning (e.g., references to chemical substances, proteins, genes, etc). The "Annotate Your Model" service uses a special format to store the annotations in data tables. This format is used in the MIRIAM.csv table file containing the annotations of model files. We also propose to use it directly within existing omics data files.

     We assume that the rows of a table refer to biological elements (e.g., metabolites, proteins, genes). Annotations for these elements are listed in special table columns. These columns can either be added manually to a data table, or semi-automatically with the aid of Annotate Your Model. Two types of such columns exist:

     1. A MiriamID column contains annotations of the same MIRIAM data type. There can be at most one annotation per element and no bioqualifiers. It is assumed that the column item and the referenced ID are linked by an "is" relationship. A table can contain several MiriamID columns, which must refer to different data resources.

     
     

     Compound	MiriamID urn:miriam:obo.chebi	MiriamID urn:miriam:kegg.compound	...
     water	CHEBI:15377	C00001	...
     ATP	CHEBI:15422	C00002	...
     phosphate	CHEBI:18367		...

     
     

     To translate a column element into a valid MIRIAM URN, the Miriam URI in the header (e.g., urn:miriam:obo.chebi) is concatenated with the column item with colon used as a separator. The column item has to be translated into a URN-encoded form (see URN spec). For instance, the colon in CHEBI:16865 has to be replaced by the string "%3A" to create the URN urn:miriam:obo.chebi:CHEBI%3A16865.

     2. More complex annotations, including bioqualifiers and multiple annotations involving the same MIRIAM data type can be stored in a column called MiriamAnnotations. In this case, the entire URN-encoded URN is stored in the column item. Each column item contains a json string listing all annotations; each annotation consists of the bioqualifier and the MIRIAM URN. If no qualifier is given, the qualifier "unknown" will be assumed by default.

     
     

     Compound	MiriamAnnotations	...
     water	[["urn:miriam:obo.chebi:CHEBI%3A15377"],["urn:miriam:kegg.compound:C00001"]]	...
     ATP	[["bqbiol:hasPart","urn:miriam:obo.chebi:CHEBI%3A18367"]]	...

     
     
     

   • search for similar models in BioModels Database (redirection to our service at semanticSBML 2.0)
   • view the hightlighted source code of your model. The highlighing includes direct hyperlinks to the annotated resources. This feature exists only for model files, not for data tables.