Optimization and Prediction Models Overview

OpenProtein.AI’s Optimization and Prediction Models (OP Models) are a set of machine learning tools designed to analyze and optimize sequences that have measured properties. If you don’t have experimental data, use PoET to generate sequences. If you do have experimental data, create a project for each specific protein sequence you’re interested in exploring. OP Models allows you to:

• Visualize your data using different graphical formats.
• Use Models to train machine learning models and design sequences within your specified constraints.
• Use Substitution Analysis to make property predictions for a sequence and all of its single substitution variants.
• Use the Design tool to customize objectives for each project.

You can create unlimited projects to house different protein sequences.

To create a project, you will need a dataset formatted as a CSV table with columns containing the full sequence of each variant and their measurement values. Missing measurements are acceptable.

On the OpenProtein.AI home page, select New Project.

Name your project, then select Create.

Congratulations! You’re ready to upload your dataset.