Specificity-driven protein binder design with Odin-Multi
A useful protein binder is defined as much by what it does not bind as by what it does. Some applications call for one binder to cover a family of related targets; others require it to distinguish a single member from near-identical relatives. Yet, widely used deep-learning-based de novo design methods typically optimise one interaction at a time, leaving cross-reactivity and specificity to…
Protein binders are characterized by their ability to avoid binding to unwanted targets while selectively interacting with desired ones. In de novo design methods, deep learning is often used to optimize individual interactions, leaving cross-reactivity and specificity to emerge later in screening. Odin-Multi is a binder design framework that addresses this issue by optimizing a shared binder sequence against multiple complexes at once.
This shared optimization approach applies attractive objectives to target proteins and repulsive objectives to off-target proteins.
Odin-Multi was tested in silico across three systems with varying cross-reactivity and specificity challenges: class B1 G protein-coupled receptors (GPCRs), short-chain three-finger toxins, and peptide-MHC (pMHC) complexes. For pairs of related GPCRs, 83.5 to 96.8% of jointly optimized designs surpassed an interaction-confidence threshold for both targets, compared to only 6.8 to 36.3% for designs optimized against a single target.
In the case of two short-chain three-finger neurotoxins, 9.2% of jointly optimized designs met the threshold for both targets, while designs optimized against one toxin alone achieved only 0.8%. In a pMHC specificity benchmark, Odin-Multi increased the fraction of designs meeting target-confidence and target-to-off-target interaction-confidence ratio of 2.5 from 6.0% to 14.2%.
Experimental screening of Odin-Multi-designed leads in both systems confirmed the computational predictions. Notably, a cross-reactive toxin minibinder showed apparent nanomolar binding to both the neurotoxin Erabutoxin A and a candidate NK-shNTx-containing fraction from Naja kaouthia venom, with higher-affinity fitted components of 11.95 and 34.43 nM, respectively.
Additionally, a pMHC minibinder demonstrated superior target-to-off-target discrimination compared to a previously reported design. By treating cross-reactivity and specificity as explicit design objectives, Odin-Multi expands the range of binding behaviors accessible to computational design.
Written by urgent.news from bioRxiv's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
