Overview

BioTransformer Options

The user can select between 8 different options for metabolism prediction or metabolite identification. The first 5 options allow for a specific coverage of metabolism and are CYP450, EC-Based, Phase II, Gut Microbial, and Environmental Microbial options. Two options allow for a comprehensive coverage of metabolism in the human superorganism and include the allHuman and SuperBio options. The last option is Multi-Step which allows a user to specify the order of individual options.

  • CYP450: Select this option for CYP450 metabolism prediction. Cytochromes P450 are a superfamily of enzymes that play a very important role in drug metabolism and usually increases the water solubility of the substrate, converting lipid-soluble compounds to more water-soluble ones, often through oxidation. These enzymes metabolize thousands of endogenous and exogenous compounds. Some metabolize one (or only a few) substrates while others may metabolize multiple substrates. CYP450 metabolism involves most of the reactions in Phase I metabolism, whereby a compound’s reactivity is increased and hydrophilicity is slightly increased.
  • EC-Based: EC-Based: Select this option for the prediction of promiscuous metabolism (e.g., glycerolipid metabolism). The Enzyme Commission (EC) assigns numbers to enzymes based on the reactions they catalyze. These enzymes may have multiple substrates, products, and/or reactions, and hence the lack of strict specificity makes them “promiscuous”. This prediction option generates metabolites using reaction rules associated with different enzymes based on the EC numbering scheme.
  • Phase II: Select this option for the prediction of major conjugative reactions, including glucuronidation, sulfation, glycine transfer, N-acetyl transfer, and glutathione transfer, amongst others. Phase II reactions strongly increase hydrophilicity to detoxify compounds by enabling renal excretion.
  • Gut Microbial: Select this option to predict xenobiotic (i.e., exogenous compounds) metabolism by gut microbial enzymes. The gut microbiota is very important in transforming dietary components, drugs, and industrial chemicals into metabolites, and its chemistry is often distinct from that of the host’s enzymes. Microbiotic enzymatic reactions are typically hydrolytic or reductive in nature, in contrast to host enzymatic reactions being oxidative or conjugative.
  • Environmental microbial: Select this option to predict metabolism or degradation of small organic molecules by environmental microbes.
  • AllHuman: Select this option to predict small molecule metabolism in the human superorganism. This covers biotransformations occurring both in human tissues as well as the gut microbiota. Each step of the metabolism prediction covers human as well as gut microbial transformations, when applicable.
  • SuperBio: SuperBio was originally designed to explore all possible metabolites in the human body for the query compound by running the allHuman module till there are no more novel metabolites generated. However, the SuperBio module currently on our web server runs the allHuman module for only 4 iterations in order to save time and conserve resources so that BioTransformer is available to run queries from more users. To run the original SuperBio module, please download the jar file and run it locally.
  • MultiBio: Select this option to create your own biotransformation sequence via any combination of CYP450, EC-Based, Phase II, Gut Microbial, and Environmental Microbial metabolism options. You may choose up to 3 iterations for each option, subject to an overall total of 4.
  • AbioticBio: Select this option to make predictions for abiotic transformations including photochemical, chlorination, ozonation reaions, etc.

Web Server Accessibility

The webserver can be freely accessed in two ways: Users can submit queries, and retrieve results both manually and programmatically as described on the next pages.

Fair usage policy

In order to ensure that resources are shared properly, we have limited the number of starting compounds per query to 1. Thus, the tab-separated structural input should have no more than 1 line, and the SD File (SDF) must contain no more than 1 compound. Users who would like to submit more compounds at once or predict more multi-step metabolism are asked to used the command-line executable available here.

Metabolism Prediction

The Input

In order to predict metabolism, the BioTransformer Metabolism Prediction Tool (BMPT) accepts five parameters as input, described as follows:
  • The task type (required): The user must select the “Metabolism Prediction” task.
  • The BioTransformer option (required): The user can select one of the options depending of the aspect of metabolism that is of interest. The prediction could be specific (CYP450, Phase II, or EC-based), or comprehensive (Gut microbial, Environmental microbial, AllHuman, or Superbio). For a detailed explanation of the different option, please read the Overview page.
  • The structure input (required): The input structures for starting compounds can be submitted in two different ways:
    • A tab-separated text: Each line contains either a (preferably isomeric) SMILESstring, or a standard InChI of a starting compound. The structural representation (SMILES or standard InChI) can be preceded by an identifier (e.g. a name, a database ID).
    • An SDF file: The SDF can be uploaded, and must contain the exact structure of each starting compound. For each starting compound, the optional identifier should be inserted in the header, just before the corresponding connection table. For more information, please consult the following document.
  • The Number of iterations (optional): This determines the maximal number of biotransformation steps that will be predicted.
  • The Submission title (optional): This allows the user to specify a title for the query. Please keep in mind that queries are accessible from the web server only through their query ID, as returned by BioTransformer.

The Output

As default, the query output is returned in an HTML format that is displayed on the web interface.

Moreover, they can download the results in a single JSON, SDF, or CSV file.

Metabolite Identification

The users now can identify the metabolites of their interest using the Chemical Formula, Min Mass and max Mass filters. For example, if the user wants to identify the metabolites whose mass equals 208.0735 with a tolerance of 0.005 for Epicatechin, they should fill 208.0685 in the Min Mass filter and 208.0785 in the Max Mass filter. The result page will then only present the predicted metabolites that satisfy the conditions.

As default, the query output is presented in an HTML format that is displayed on the web interface.
Example of using filters to identify molecuels on the prediction result html page.

Programmatic Access

Resources

There is one end-point available to submit queries and retrieve the results from the queries performed:
biotransformer.ca/queries.json

Submit query (POST)

POST: Submit a query to the biotransformer app. In the production server, the number of post requests is limited to 2 per minute. The attributes are listed below. Only one of the parameters ‘query_input’ or ‘fstruc’ must be provided with content.

query_label (optional): any string for labeling the query. task_type (required): ‘PREDICTION’. biotransformer_option (required): ‘CYP450’, ‘EC-BASED’, ‘PHASEII’, ‘HGUT’, ‘ENVMICRO’, ‘ALLHUMAN’, ‘SUPERBIO’, 'MULTIBIO', 'ABIOTICBIO'.
query_input (optional): this option is used only if a the user wants to enter a text input, as described in the ‘Metabolism Prediction’ section. The tab spaces must be replaced by a tab (‘\t’) or space ('\s') character, and the newlines must be replaced by ‘\n’. fstruct (optional): The path to the SD File to be uploaded. number_of_steps (optional): It determines the maximal number of biotransformation steps that will be predicted. (http://biotransformer.ca/).
This is the command that can be used with curl to use the REST API:

curl -i -H "Content-Type: application/json" -H "Accept: application/json" http://biotransformer.ca/queries.json -X POST -d '{"biotransformer_option":{OPTION}, "number_of_steps":1, "query_input":{QUERY_INPUT} ,"task_type":{TASK TYPE}}'

Examples of query submissions are shown below:

curl -i -H "Content-type: application/json" -H "Accept: application/json" http://biotransformer.ca/queries.json -X POST -d '{"biotransformer_option":"CYP450", "number_of_steps":1, "query_input":"acetaminophen\tCC(=O)NC1=CC=C(O)C=C1", "task_type":"PREDICTION" }'

Where the json sent is:

        

If the number of allowed requests is reached in the production server, the user will see the next message:
Fig. 1: post requests limit reached in the production server
Fig. 1: post requests limit reached in the production server.

Retrieve results (GET)

curl -H "Accept: application/json" -X GET http://biotransformer.ca/queries/{QUERY_NUMBER}.json The query number can be shown in the response from the server when performing a POST request (see image below):
Fig. 2: example of a POST request. Query id remarked
Fig. 2: example of a POST request. Query id remarked.



The attributes of the query results for predictions are:
    id: id of the query. label: null or string specified by input query. task: ‘PREDICTION’, ‘IDENTIFICATION’. biotransformer_option: ‘ENVMICRO’, ‘ALLHUMAN’, ‘SUPERBIO’. number_of_steps: number of steps to run. status: ‘failed’, ‘In progress’, ‘Done’. number_of_starting_compounds: number of compounds in the entry. Limited to 1. total_prediction_time_in_ms: number of ms used for processing the biotransformation. "invalid_compounds": array of invalid compounds introduced by the user "predictions": results of the metabolites predicted, products of the biotransformer processing. "number_of_unique_metabolites": number of unique metabolites. "prediction_errors": errors ocurred during the prediction
A json example of the query results is shown below:

    

Browser Compliance

OS Version Chrome Firefox Microsoft Edge Safari
Linux Ubuntu 18.04.1 lts 96.0.4464.110-1 95.02 96.0.1054.62 N/A
MacOS Monterey 12.0.1 96.0.4464.93 95.02 96.0.1054.62 15.1
Windows 10 96.0.4464.110.1 95.02 96.0.1054.62 N/A

Substructure Filter

The users can input the SMILES String of the substructure they want to search. The result page will show all the metabolite structures that contain the query substrucure among all the predicted results.

Example

The user wants to know which metabolites produced by AllHuman metabolism for glucose contains phosphates. He should do following steps: Figure out and create the SMILES String for phosphate, which is OP(O)(O)=O Input the SMILES String into the context field uner the Filter by Substructure Click the search button The result page will reload and present the metabolites that contain any number of phosphate groups Please delete the SMILES and click the search button again to get the original full predicted results