iiinova — where your digital self keeps evolving

Ben Brown @bpbrown.bsky.social · 02/04/2026

Super proud of Katy Butler for being awarded a Goldwater Scholarship! news.ua.edu/2026/04/3-ua.... Katy interned in my lab, where she worked on formulating protein conformational sampling on quantum annealers. She is exceptional - look out for her PhD!

news.ua.edu

3 UA Students Awarded Prestigious Goldwater Scholarships

The Barry Goldwater Scholarship and Excellence in Education Program has selected three UA students as Goldwater Scholars for 2026-2027.

010

Ben Brown @bpbrown.bsky.social · 18/10/2025

Good question. Yeah, I had reviews from two folks and also feedback from the editor. I do not know the rules or customs around anonymity of reviewers at PNAS. Most journals from which I receive feedback do not disclose reviewer identities. There is some discussion around this practice broadly.

010

Ben Brown @bpbrown.bsky.social · 17/10/2025

Most of the posts I see on peer review during the publication process are overwhelmingly negative. I just wanted to highlight that in this case I received helpful feedback. I have preprinted before and will likely do so again in the future.

010

Ben Brown @bpbrown.bsky.social · 17/10/2025

Big thanks for all the support from the new Vanderbilt Center for AI in Protein Dynamics, our CSB @vanderbiltcsb.bsky.social , and my department.

000

Ben Brown @bpbrown.bsky.social · 17/10/2025

I started my lab in April 2024, and I think we are starting to build some momentum. Hopefully in the next few months I will be able to share some other stuff we are working on.

100

Ben Brown @bpbrown.bsky.social · 17/10/2025

Finally, I want to note that the peer review process improved this manuscript. While I understand the benefits of pre-prints and the limitations of peer review, this was an instance where it was genuinely constructive and elevated the final paper.

311

Ben Brown @bpbrown.bsky.social · 17/10/2025

Also, apologies for being slow with it, but I'll add more scripts, examples, a better UI, etc. to the GitHub soon. I had to freeze a lot of the project a while back for purposes of benchmarks and picking a stopping point for the manuscript.

100

Ben Brown @bpbrown.bsky.social · 17/10/2025

The new CORDIAL model will also be trained on substantially more synthetic null data covering broader chemical and structural perturbations. We will upload the weights for these new models, too.

100

Ben Brown @bpbrown.bsky.social · 17/10/2025

Co-folding models are likely overtrained on pairs of sequences and chemical substructures, but for generating plausible structures for affinity prediction with CORDIAL, they probably represent a better version of what I tried to do with the MCS maps and refinement.

100

Ben Brown @bpbrown.bsky.social · 17/10/2025

Speaking of improvements, we're extending the training set with the new SAIR dataset from @sandboxaq.bsky.social . Our original augmentation mimicked known poses via MCS mapping.

100

Ben Brown @bpbrown.bsky.social · 17/10/2025

Anyway, CORDIAL generalized pretty well, but perhaps not unexpectedly it did not yield dramatic performance improvements over Vina. There are clear ways to increase CORDIAL's expressivity - learning atom-pair embeddings/weights, incorporating additional geometric information, etc.

100

Ben Brown @bpbrown.bsky.social · 17/10/2025

This first pass at the CATH-LSO benchmark was useful, but in subsequent iterations I'll be tweaking it to make it more challenging. I hope this work encourages more dialogue on best practices for retrospective validation. I'm open to better strategies if folks have suggestions.

100

Ben Brown @bpbrown.bsky.social · 17/10/2025

So, how do we evaluate generalizability? I tried to set it up to mimic screening against a member of a novel, unseen protein superfamily. I hold out a protein superfamily and its associated chemistry, train on the remainder, and test on the held-out set. Thanks CATH team.

100

Ben Brown @bpbrown.bsky.social · 17/10/2025

This doesn't completely eliminate bias, but it reduces it and makes it more predictable. For example, signal magnitudes in feature columns can differ between train/test sets. Consequently, BatchNorm1d or something similar is required to prevent the model from over-training on these patterns.

100

Ben Brown @bpbrown.bsky.social · 17/10/2025

The approach here was to use a task-specific architecture. Instead of guiding the model to focus on interactions, we restrict its learning space to them. The model is constrained to view the problem only through distance-dependent physicochemical pairings.

100

Ben Brown @bpbrown.bsky.social · 17/10/2025

The challenge is that the model needs a massive amount of data to guide it to learning the problem how we want. With a broad inductive bias, a model can easily learn non-causal correlations from training set artifacts instead of the generalizable principles we intend.

100

Ben Brown @bpbrown.bsky.social · 17/10/2025

Often we have an idea of what we want the model to learn, and it is easy to assume that the network will tend to learn the problem the way that we consider it.

100

Ben Brown @bpbrown.bsky.social · 17/10/2025

This manuscript is an exploration of learning spaces. In my lab, we think a lot about the spaces of things. A model's architecture defines the manifold on which learning occurs.

100

Ben Brown @bpbrown.bsky.social · 17/10/2025

To be clear, I am not selling a model, I do not believe I have solved this problem, and I am not suggesting you should scrap your existing tools and just use this. The paper introduces a model, CORDIAL, but it's not really about the model itself. So, what is this manuscript about?

110

Ben Brown @bpbrown.bsky.social · 17/10/2025

My first manuscript as an independent PI, and my first single-author research article, is now published in @pnas.org. It's an attempt to contribute to the dialogue on generalizability in structure-based protein-small molecule affinity prediction with neural networks.

1113

Reposted by Ben Brown

Frank Noe @franknoe.bsky.social · 06/12/2024

Super excited to preprint our work on developing a Biomolecular Emulator (BioEmu): Scalable emulation of protein equilibrium ensembles with generative deep learning from @msftresearch.bsky.social ch AI for Science. www.biorxiv.org/content/10.1...

21441147