Pareto Optimization of Combinatorial Mutagenesis Libraries

Machine learning-guided co-optimization of fitness and diversity facilitates combinatorial library design in enzyme engineering

The effective design of combinatorial libraries to balance fitness and diversity facilitates the engineering of useful enzyme functions, particularly those that are poorly characterized or unknown in biology. We introduce MODIFY, a machine learning (ML) algorithm that learns from natural protein sequences to infer evolutionarily plausible mutations and predict enzyme fitness. MODIFY co-optimizes predicted fitness and sequence diversity of starting libraries, prioritizing high-fitness variants while ensuring broad sequence coverage. In silico evaluation shows that MODIFY outperforms state-of-the-art unsupervised methods in zero-shot fitness prediction and enables ML-guided directed evolution with enhanced efficiency. Using MODIFY, we engineer generalist biocatalysts derived from a thermostable cytochrome c to achieve enantioselective C-B and C-Si bond formation via a new-to-nature carbene transfer mechanism, leading to biocatalysts six mutations away from previously developed enzymes while exhibiting superior or comparable activities. These results demonstrate MODIFY's potential in solving challenging enzyme engineering problems beyond the reach of classic directed evolution.

Optimal trade-off control in machine learning-based library design, with application to adeno-associated virus (AAV) for gene therapy

Adeno-associated viruses (AAVs) hold tremendous promise as delivery vectors for gene therapies. AAVs have been successfully engineered-for instance, for more efficient and/or cell-specific delivery to numerous tissues-by creating large, diverse starting libraries and selecting for desired properties. However, these starting libraries often contain a high proportion of variants unable to assemble or package their genomes, a prerequisite for any gene delivery goal. Here, we present and showcase a machine learning (ML) method for designing AAV peptide insertion libraries that achieve fivefold higher packaging fitness than the standard NNK library with negligible reduction in diversity. To demonstrate our ML-designed library's utility for downstream engineering goals, we show that it yields approximately 10-fold more successful variants than the NNK library after selection for infection of human brain tissue, leading to a promising glial-specific variant. Moreover, our design approach can be applied to other types of libraries for AAV and beyond.

DeCOIL: Optimization of Degenerate Codon Libraries for Machine Learning-Assisted Protein Engineering

With advances in machine learning (ML)-assisted protein engineering, models based on data, biophysics, and natural evolution are being used to propose informed libraries of protein variants to explore. Synthesizing these libraries for experimental screens is a major bottleneck, as the cost of obtaining large numbers of exact gene sequences is often prohibitive. Degenerate codon (DC) libraries are a cost-effective alternative for generating combinatorial mutagenesis libraries where mutations are targeted to a handful of amino acid sites. However, existing computational methods to optimize DC libraries to include desired protein variants are not well suited to design libraries for ML-assisted protein engineering. To address these drawbacks, we present DEgenerate Codon Optimization for Informed Libraries (DeCOIL), a generalized method that directly optimizes DC libraries to be useful for protein engineering: to sample protein variants that are likely to have both high fitness and high diversity in the sequence search space. Using computational simulations and wet-lab experiments, we demonstrate that DeCOIL is effective across two specific case studies, with the potential to be applied to many other use cases. DeCOIL offers several advantages over existing methods, as it is direct, easy to use, generalizable, and scalable. With accompanying software (https://github.com/jsunn-y/DeCOIL), DeCOIL can be readily implemented to generate desired informed libraries.

Problem of Domain/Building Block Preservation in the Evolution of Biological Macromolecules and Evolutionary Computation

Structurally and functionally isolated domains in biological macromolecular evolution, both natural and artificial, are largely similar to "schemata", building blocks (BBs), in evolutionary computation (EC). The problem of preserving in subsequent evolutionary searches the already found domains / BBs is well known and quite relevant in biology as well as in EC. Both biology and EC are seeing parallel and independent development of several approaches to identifying and preserving previously identified domains / BBs. First, we notice the similarity of DNA shuffling methods in synthetic biology and multi-parent recombination algorithms in EC. Furthermore, approaches to computer identification of domains in proteins that are being developed in biology can be aligned with BB identification methods in EC. Finally, approaches to chimeric protein libraries optimization in biology can be compared to evolutionary search methods based on probabilistic models in EC. We propose to validate the prospects of mutual exchange of ideas and transfer of algorithms and approaches between evolutionary systems biology and EC in these three principal directions. A crucial aim of this transfer is the design of new advanced experimental techniques capable of solving more complex problems of in vitro evolution.

Peptide design by optimization on a data-parameterized protein interaction landscape

Many applications in protein engineering require optimizing multiple protein properties simultaneously, such as binding one target but not others or binding a target while maintaining stability. Such multistate design problems require navigating a high-dimensional space to find proteins with desired characteristics. A model that relates protein sequence to functional attributes can guide design to solutions that would be hard to discover via screening. In this work, we measured thousands of protein-peptide binding affinities with the high-throughput interaction assay amped SORTCERY and used the data to parameterize a model of the alpha-helical peptide-binding landscape for three members of the Bcl-2 family of proteins: Bcl-x_L, Mcl-1, and Bfl-1. We applied optimization protocols to explore extremes in this landscape to discover peptides with desired interaction profiles. Computational design generated 36 peptides, all of which bound with high affinity and specificity to just one of Bcl-x_L, Mcl-1, or Bfl-1, as intended. We designed additional peptides that bound selectively to two out of three of these proteins. The designed peptides were dissimilar to known Bcl-2-binding peptides, and high-resolution crystal structures confirmed that they engaged their targets as expected. Excellent results on this challenging problem demonstrate the power of a landscape modeling approach, and the designed peptides have potential uses as diagnostic tools or cancer therapeutics.