Try it and let us know how it goes!

If you want to integrate it in your own AF3 installation, check the 3 files which only need a slight modification (model.py, confidences.py and confidence_types.py):
github.com/Kuhlman-Lab/...

GitHub - Kuhlman-Lab/alphafold3: Kuhlman Lab Installation of AlphaFold3 Kuhlman Lab Installation of AlphaFold3. Contribute to Kuhlman-Lab/alphafold3 development by creating an account on GitHub.

We used the Kuhlman lab’s version of AF3, as this also makes the mmseqs2 server available for generating MSA (thank you Brian Kuhlman and Henry Dieckhaus):
github.com/Kuhlman-Lab/...

No flags required, actifptm is part of the output (also calculated between chains).

Predicting Experimental Success in De Novo Binder Design: A Meta-Analysis of 3,766 Experimentally Characterised Binders Designing high-affinity de novo protein binders has become increasingly tractable, yet in vitro prioritisation continues to depend on heuristics in the absence of systematic analysis. Here, we present...

🚨 actifpTM is now in AF3!

Since our new metric, actifpTM, was determined to be the best for predicting success in designed proteins refolded by AF2 (www.biorxiv.org/content/10.1... by @tpjenkins.bsky.social
), we made it available also in AF3!