🎬Follow-Along Platform Tour

Follow the steps shown in this video to upload your first files, create your first workspace with an AI pipeline, then turn it into a point and click tool with Ontologic.

How to use this Tutorial

First, sign up for your Ontologic trial or individual account with the link provided to you. Then you can watch this video and follow along!

Look out for the indicated time jumps where you'll have to wait for your code to run or workspace to save.

We've also included downloadable or copyable code and files for every step of the video, below.

Data

For uploading:

Boltz-2 GitHub link for background:

Not found

Workspaces Code for Copying

To update dependencies and install Boltz-2:

sudo apt update
sudp apt install python3-pip
pip install boltz[cuda] -U

Contents of affinity.yaml input file:

version: 1  # Optional, defaults to 1
sequences:
  - protein:
      id: A
      sequence: MVTPEGNVSLVDESLLVGVTDEDRAVRSAHQFYERLIGLWAPAVMEAAHELGVFAALAEAPADSGELARRLDCDARAMRVLLDALYAYDVIDRIHDTNGFRYLLSAEARECLLPGTLFSLVGKFMHDINVAWPAWRNLAEVVRHGARDTSGAESPNGIAQEDYESLVGGINFWAPPIVTTLSRKLRASGRSGDATASVLDVGCGTGLYSQLLLREFPRWTATGLDVERIATLANAQALRLGVEERFATRAGDFWRGGWGTGYDLVLFANIFHLQTPASAVRLMRHAAACLAPDGLVAVVDQIVDADREPKTPQDRFALLFAASMTNTGGGDAYTFQEYEEWFTAAGLQRIETLDTPMHRILLARRATEPSAVPEGQASENLYFQ
      msa: empty
  - ligand:
      id: B
      smiles: 'N[C@@H](Cc1ccc(O)cc1)C(=O)O'
properties:
  - affinity:
      binder: B

To run the Boltz-2 affinity prediction:

boltz predict files/affinity.yaml --out_dir ./results --num_workers 0

Tools Code for Copying

Command to set the tool:

boltz predict affinity.yaml \
   --num_workers 0 \
   --output_format pdb \
   --out_dir ./outputs 

Last updated