Optibrium Community

Downloads

Phenaris Transporter Models

Apr 29, 2021

Phenaris Transporter Models

Apr 29, 2021

This script provides the ability to run Phenaris’ transporter models and return the results directly into StarDropTo use this tool, you will need login credentials which are supplied by Phenaris. To obtain this, please request access on this website. 

Predictions can be made for 9 transporterswhich are involved in drug/drug interactions and hepatoxicity. These include prediction of inhibitors of P-glycoprotein, BCRP, BSEP, MRP3, OATP1B1, OATP1B3, OCT1, OCT2, and MATE1.  

To learn more about these transporter models, please visit this webpage and watch our joint Webinar with Phenaris.

Please note that any compound structures used in a query will be passed to the Phenaris servers. These data will be sent using a secure connection (i.e. transmitted via HTTPS) and you can view the Phenaris privacy policy.

Version

Please note: This script is compatible with StarDrop 6.6, or more recent versions, for either Windows® or Mac®. To find out which version you have installed, start StarDrop and select the Help->About menu. If you are using an older version of StarDrop please contact stardrop-support@optibrium.com and we will be happy to provide download details.

Windows

To install this script on Windows:

  • Ensure that you have saved your work and close StarDrop
  • For 64-bit Windows, download the file StarDrop Script – Phenaris.exe
  • When the download is complete, double-click the file to run it and follow the instructions (we recommend that you accept the default options provided)
  • Restart StarDrop

macOS

To install this script on macOS:

  • Ensure that you have saved your work and close StarDrop
  • Download the file StarDrop Script – Phenaris.zip
  • In your user area, navigate to the StarDrop folder (~/StarDrop/py)
  • Extract the contents of the zip file into this folder overwriting any existing files
  • Restart StarDrop

To use this script…

Download documentation on using this script for more details.

Using the Phenaris transporter models

The Plugin of Phenaris provides classification models which predict inhibitor molecules for transporters considered relevant by FDA, EMA and the Japanese regulatory agency. It is available via Custom Scripts. It is possible to predict a single molecule by using the so-called Single Prediction as well as a batch upload (Prediction) via the StarDrop upload function. When you start the plugin, you are asked to enter your credentials. You can request access via our homepage.

After login you can choose between the following three options:
• „Draw molecule for single prediction…”
• „Single prediction using selected model…“
• “Prediction…”

1., „Draw molecule for single prediction…”
• Select Custom Scripts > Phenaris > Draw molecule for single prediction
• Draw your molecule in the Structure Sketcher

At the moment there are eight classification models of transporters available namely BCRP, BSEP, MATE1, MRP3, OATP1B1, OATP1B3, OCT1, OCT2 and P-gp (MDR1). The models are based on the Random Forest Classifier. You can select one or more transporter models from the list or select all at once.

The Output of the prediction provides the classification “yes” and “no” and the prediction score. Further, the applicability domain is given as Boolean result. The results are shown in a newly created dataset called Single prediction. If multiple single predictions are made, the newly predicted molecule will be appended to the existing Single prediction dataset. If you change the model selection between the different Single Prediction runs the fields without a result obtain the entry “?” which indicates that there is no result available.

“Yes” indicates that the compound is predicted as an inhibitor and “no” indicates that the compound is predicted as a non-inhibitor. The prediction score gives the probability if a molecule will inhibit the transporter protein. The score ranges from 0.5 to 1, while a value close to 1 indicate a high probability for the prediction. A value between 0.5 and 0.65 indicate a grey area, where the result has to be taken with caution. The Applicability Domain defines the chemical space of each model. The algorithm is based on the local outlier factor (LOF) approach, which takes the local density into account. True means the test compound is within the chemical space of the model and is therefore In Domain. False mean that the test compound is Out of Domain and cannot be predicted with the given model.

2., „Single prediction using selected model…“
• You can select a compound from an existing dataset in StarDrop
• Select Custom Scripts > Phenaris > Single prediction using selected model…
• You can select transporter models for the prediction as mentioned in the section „Draw molecule for single prediction…”
• The result is appended to the dataset used for the prediction.
• A more detailed description of the results is mentioned in the section „Draw molecule for single prediction…”

3., “Prediction…”
• You first have to upload a file (e.g. SDF) or you can use an existing dataset in StarDrop
• Select Custom Scripts > Phenaris > Prediction
• You can select transporter models for the prediction as mentioned in the section „Draw molecule for single prediction…”
• The result is appended to the dataset used for the prediction.
• A more detailed description of the results is mentioned in the section „Draw molecule for single prediction…”