Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
167 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
42 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Tree Search-Based Evolutionary Bandits for Protein Sequence Optimization (2401.06173v1)

Published 8 Jan 2024 in q-bio.BM and cs.LG

Abstract: While modern biotechnologies allow synthesizing new proteins and function measurements at scale, efficiently exploring a protein sequence space and engineering it remains a daunting task due to the vast sequence space of any given protein. Protein engineering is typically conducted through an iterative process of adding mutations to the wild-type or lead sequences, recombination of mutations, and running new rounds of screening. To enhance the efficiency of such a process, we propose a tree search-based bandit learning method, which expands a tree starting from the initial sequence with the guidance of a bandit machine learning model. Under simplified assumptions and a Gaussian Process prior, we provide theoretical analysis and a Bayesian regret bound, demonstrating that the combination of local search and bandit learning method can efficiently discover a near-optimal design. The full algorithm is compatible with a suite of randomized tree search heuristics, machine learning models, pre-trained embeddings, and bandit techniques. We test various instances of the algorithm across benchmark protein datasets using simulated screens. Experiment results demonstrate that the algorithm is both sample-efficient and able to find top designs using reasonably small mutation counts.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (56)
  1. Improved algorithms for linear stochastic bandits. Advances in neural information processing systems, 24: 2312–2320.
  2. Model-based reinforcement learning for biological sequence design. In International Conference on Learning Representations.
  3. Arnold, F. H. 1998. Design by directed evolution. Accounts of chemical research, 31(3): 125–131.
  4. Auer, P. 2002. Using Confidence Bounds for Exploitation-Exploration Trade-offs. Journal of Machine Learning Research, 3: 397–422.
  5. Machine learning to design integral membrane channelrhodopsins for efficient eukaryotic expression and plasma membrane localization. PLoS computational biology, 13(10): e1005786.
  6. Conditioning by adaptive sampling for robust design. In International conference on machine learning, 773–782. PMLR.
  7. Design by adaptive sampling. arXiv preprint arXiv:1810.03714.
  8. Deep diversification of an AAV capsid protein by machine learning. Nature Biotechnology, 39(6): 691–696.
  9. Enzyme engineering for nonaqueous solvents: random mutagenesis to enhance activity of subtilisin E in polar organic media. Bio/Technology, 9(11): 1073–1077.
  10. Tuning the activity of an enzyme for unusual environments: sequential random mutagenesis of subtilisin E for catalysis in dimethylformamide. Proceedings of the National Academy of Sciences, 90(12): 5618–5622.
  11. On kernelized multi-armed bandits. In International Conference on Machine Learning, 844–853. PMLR.
  12. Doppa, J. R. 2021. Adaptive experimental design for optimizing combinatorial structures. In Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence (IJCAI), 4940–4945.
  13. ProtTrans: Towards Cracking the Language of Life’s Code Through Self-Supervised Deep Learning and High Performance Computing.
  14. A dataset comprised of binding interactions for 104,972 antibodies against a SARS-CoV-2 peptide. Scientific Data, 9(1): 653.
  15. Autofocused oracles for model-based design. Advances in Neural Information Processing Systems, 33: 12945–12956.
  16. Optimizing the search algorithm for protein engineering by directed evolution. Protein engineering, 16(8): 589–597.
  17. Improving catalytic function by ProSAR-driven enzyme evolution. Nature biotechnology, 25(3): 338–344.
  18. Machine learning to navigate fitness landscapes for protein engineering. Current Opinion in Biotechnology, 75: 102713.
  19. Pervasive pairwise intragenic epistasis among sequential mutations in TEM-1 β𝛽\betaitalic_β-lactamase. J. Mol. Biol., 431(10): 1981–1992.
  20. High-dimensional sparse linear bandits. Advances in Neural Information Processing Systems, 33: 10753–10763.
  21. Directed evolution strategies for improved enzymatic performance. Microbial Cell Factories, 4(1): 1–6.
  22. Learning protein fitness models from evolutionary and assay-labeled data. Nature Biotechnology, 40(7): 1114–1122.
  23. Neural tangent kernel: Convergence and generalization in neural networks. Advances in neural information processing systems, 31.
  24. Directed evolution of enzyme catalysts. Trends in biotechnology, 15(12): 523–530.
  25. Rarely-switching linear bandits: optimization of causal effects for the real world. arXiv preprint arXiv:1905.13121.
  26. Communication efficient distributed learning for kernelized contextual bandits. Advances in Neural Information Processing Systems, 35: 19773–19785.
  27. A contextual-bandit approach to personalized news article recommendation. In Proceedings of the 19th international conference on World wide web, 661–670. ACM.
  28. Mockus, J. 1989. Bayesian Approach to Global Optimization: Theory and Applications.
  29. ProGen2: Exploring the Boundaries of Protein Language Models.
  30. Methods for the directed evolution of proteins. Nature Reviews Genetics, 16(7): 379–394.
  31. Evaluating Protein Transfer Learning with TAPE. In Advances in Neural Information Processing Systems.
  32. Proximal exploration for model-guided protein sequence design. In International Conference on Machine Learning, 18520–18536. PMLR.
  33. Biological Structure and Function Emerge from Scaling Unsupervised Learning to 250 Million Protein Sequences. PNAS.
  34. Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences. Proceedings of the National Academy of Sciences, 118(15): e2016239118.
  35. Navigating the protein fitness landscape with Gaussian processes. Proceedings of the National Academy of Sciences, 110(3): E193–E201.
  36. Learning to optimize via posterior sampling. Mathematics of Operations Research, 39(4): 1221–1243.
  37. Seeger, M. 2004. Gaussian processes for machine learning. International journal of neural systems, 14(02): 69–106.
  38. Protein design and variant prediction using autoregressive generative models. Nature communications, 12(1): 1–11.
  39. Generative language modeling for antibody design. bioRxiv.
  40. AdaLead: A simple and robust adaptive greedy search algorithm for sequence design. arXiv preprint.
  41. Phage display. Chemical reviews, 97(2): 391–410.
  42. Gaussian process optimization in the bandit setting: No regret and experimental design. arXiv preprint arXiv:0912.3995.
  43. Accelerating Bayesian Optimization for Biological Sequence Design with Denoising Autoencoders. In Chaudhuri, K.; Jegelka, S.; Song, L.; Szepesvari, C.; Niu, G.; and Sabato, S., eds., Proceedings of the 39th International Conference on Machine Learning, volume 162 of Proceedings of Machine Learning Research, 20459–20478. PMLR.
  44. Tuned Fitness Landscapes for Benchmarking Model-Guided Protein Design. bioRxiv.
  45. Turner, N. J. 2009. Directed evolution drives the next generation of biocatalysts. Nature chemical biology, 5(8): 567–573.
  46. Neural Bandits for Protein Sequence Optimization. In 2022 56th Annual Conference on Information Sciences and Systems (CISS), 188–193. IEEE.
  47. Gaussian processes for machine learning, volume 2. MIT press Cambridge, MA.
  48. Making antibodies by phage display technology. Annual review of immunology, 12(1): 433–455.
  49. Machine learning-assisted directed protein evolution with combinatorial libraries. Proceedings of the National Academy of Sciences, 116(18): 8852–8858.
  50. Neural Contextual Bandits with Deep Representation and Shallow Exploration. CoRR, abs/2012.01780.
  51. Machine-learning-guided directed evolution for protein engineering. Nature methods, 16(8): 687–694.
  52. Sample-optimal parametric q-learning using linearly additive features. In International Conference on Machine Learning, 6995–7004. PMLR.
  53. Reinforcement learning in feature space: Matrix bandit, kernels, and regret bound. In International Conference on Machine Learning, 10746–10756. PMLR.
  54. Bandit Theory and Thompson Sampling-Guided Directed Evolution for Sequence Optimization. arXiv preprint arXiv:2206.02092.
  55. Neural thompson sampling. arXiv preprint arXiv:2010.00827.
  56. Neural contextual bandits with ucb-based exploration. In International Conference on Machine Learning, 11492–11502. PMLR.
Citations (1)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now