Apr 27, 2026
Activation Probes for Synthetic Toxin Variant Detection
Maxwell DeFanti, Ishaan Panigraphi, Kevin Zhang
DNA synthesis screening prevents bad actors from obtaining the physical sequences needed to produce dangerous toxins and pathogens. However, current screening tools like BLAST and SecureDNA rely on sequence similarity to known threats, and recent work has shown that AI protein design tools such as ProteinMPNN can generate functional toxin variants that evade these screens at rates approaching 100%. We introduce a screening approach that trains an activation probe on ESM-2 embeddings to recognize toxic function across diverged sequences; on held-out synthetic variants that are ~40% identity to their parents, our classifier maintains 86.7% recall while BLAST recall collapses to 46.7%. This provides initial evidence that protein language model embeddings can be a robust second layer of defense for DNA synthesis screening, complementing current similarity-based methods.
This addresses the right problem at the right time and the pipeline design is genuinely thoughtful, particularly the cluster-aware evaluation and the ablation showing ESM-2 embeddings carry enough functional signal to generalise without synthetic training data. The central weakness is that the headline results rest on 15 sequences per divergence level, which means the recall numbers could shift substantially with a handful of different outcomes. Acknowledging variance across runs is honest, but it also undercuts confidence in the specific numbers the paper leads with. The false positive rate tripling relative to BLAST also needs more engagement, because in a real screening deployment that’s the number that determines whether providers actually adopt the tool owing to operational costs. Scaling up the synthetic evaluation set and stress-testing the false positive rate at operationally realistic thresholds would turn this from a promising proof of concept into something deployable.
Toxins/toxicity are a really important focus area. I fully agree with the importance of more advanced synthesis screening approaches, and testing utility of ESM-2 embedding may help drive toward function-based approaches that we need. Results seem reasonable and clearly reported. If authors pursue this further, I would suggest continuing the thread on toxins as equally important to the task of applying the approach to bacteria and viruses (as they are publicly noted in terms of potential bioweapons activities by certain countries in the latest State Department treaty compliance reports). Appreciate the focus on ESM-2 rather than other open models for which potential risks/info hazards are less well explored to date.
You have generated important new knowledge and contributed to the challenge of moving from sequence-based to function-based screening. I loved the discussion on the importance of lab experimental confirmation and the challenges it brings. I would like to have understood a little but more about how the time and compute factors limited your work. I think understanding the limits of what such approaches might achieve, and connecting that to resource availability adds another dimension to this interesting and important challenge.
Cite this work
@misc {
title={
(HckPrj) Activation Probes for Synthetic Toxin Variant Detection
},
author={
Maxwell DeFanti, Ishaan Panigraphi, Kevin Zhang
},
date={
4/27/26
},
organization={Apart Research},
note={Research submission to the research sprint hosted by Apart.},
howpublished={https://apartresearch.com}
}
