AffinityFlow: Guided Flows for Antibody Affinity Maturation

Can Chen, Karla-Luise Herpoldt, Chenchao Zhao, Zichen Wang, Marcus D. Collins, Shang Shang, Ron Benson
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:8212-8225, 2025.

Abstract

Antibodies are widely used as therapeutics, but their development requires costly affinity maturation, involving iterative mutations to enhance binding affinity. This paper explores a sequence-only scenario for affinity maturation, using solely antibody and antigen sequences. Recently AlphaFlow wraps AlphaFold within flow matching to generate diverse protein structures, enabling a sequence-conditioned generative model of structure. Building on this, we propose an alternating optimization framework that (1) fixes the sequence to guide structure generation toward high binding affinity using a structure-based predictor, then (2) applies inverse folding to create sequence mutations, refined by a sequence-based predictor. A key challenge is the lack of labeled data for training both predictors. To address this, we develop a co-teaching module that incorporates valuable information from noisy biophysical energies into predictor refinement. The sequence-based predictor selects consensus samples to teach the structure-based predictor, and vice versa. Our method, AffinityFlow, achieves state-of-the-art performance in proof-of-concept affinity maturation experiments.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-chen25y, title = {{A}ffinity{F}low: Guided Flows for Antibody Affinity Maturation}, author = {Chen, Can and Herpoldt, Karla-Luise and Zhao, Chenchao and Wang, Zichen and Collins, Marcus D. and Shang, Shang and Benson, Ron}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {8212--8225}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/chen25y/chen25y.pdf}, url = {https://proceedings.mlr.press/v267/chen25y.html}, abstract = {Antibodies are widely used as therapeutics, but their development requires costly affinity maturation, involving iterative mutations to enhance binding affinity. This paper explores a sequence-only scenario for affinity maturation, using solely antibody and antigen sequences. Recently AlphaFlow wraps AlphaFold within flow matching to generate diverse protein structures, enabling a sequence-conditioned generative model of structure. Building on this, we propose an alternating optimization framework that (1) fixes the sequence to guide structure generation toward high binding affinity using a structure-based predictor, then (2) applies inverse folding to create sequence mutations, refined by a sequence-based predictor. A key challenge is the lack of labeled data for training both predictors. To address this, we develop a co-teaching module that incorporates valuable information from noisy biophysical energies into predictor refinement. The sequence-based predictor selects consensus samples to teach the structure-based predictor, and vice versa. Our method, AffinityFlow, achieves state-of-the-art performance in proof-of-concept affinity maturation experiments.} }
Endnote
%0 Conference Paper %T AffinityFlow: Guided Flows for Antibody Affinity Maturation %A Can Chen %A Karla-Luise Herpoldt %A Chenchao Zhao %A Zichen Wang %A Marcus D. Collins %A Shang Shang %A Ron Benson %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-chen25y %I PMLR %P 8212--8225 %U https://proceedings.mlr.press/v267/chen25y.html %V 267 %X Antibodies are widely used as therapeutics, but their development requires costly affinity maturation, involving iterative mutations to enhance binding affinity. This paper explores a sequence-only scenario for affinity maturation, using solely antibody and antigen sequences. Recently AlphaFlow wraps AlphaFold within flow matching to generate diverse protein structures, enabling a sequence-conditioned generative model of structure. Building on this, we propose an alternating optimization framework that (1) fixes the sequence to guide structure generation toward high binding affinity using a structure-based predictor, then (2) applies inverse folding to create sequence mutations, refined by a sequence-based predictor. A key challenge is the lack of labeled data for training both predictors. To address this, we develop a co-teaching module that incorporates valuable information from noisy biophysical energies into predictor refinement. The sequence-based predictor selects consensus samples to teach the structure-based predictor, and vice versa. Our method, AffinityFlow, achieves state-of-the-art performance in proof-of-concept affinity maturation experiments.
APA
Chen, C., Herpoldt, K., Zhao, C., Wang, Z., Collins, M.D., Shang, S. & Benson, R.. (2025). AffinityFlow: Guided Flows for Antibody Affinity Maturation. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:8212-8225 Available from https://proceedings.mlr.press/v267/chen25y.html.

Related Material