Enzyme Engineering Services
Enzyme Engineering and Directed Evolution
Evolve biocatalysts with improved activity, specificity, and stability for bio-based polymer and specialty chemical processes.
What Directed Evolution Delivers
Directed evolution applies the principles of natural selection in the laboratory. Starting from an enzyme that partially meets your requirements, we introduce genetic diversity across the encoding gene, apply selection pressure for the property you care about, and recover variants with improved performance. The process is iterative: each round of mutagenesis and selection enriches the population for better variants, and over multiple rounds the enzyme accumulates mutations that shift its properties toward your specification.
This approach is valuable precisely because it does not require prior structural knowledge. Where structure-guided or computational design depends on a reliable model, directed evolution lets selection find what works. In practice, the strongest campaigns combine the two: a model trained on your own selection data proposes where to diversify next, the laboratory round returns real fitness measurements on thousands of variants, and those measurements retrain the model. We pick the entry point honestly and match the method to the problem.
Catalytic Activity
Campaigns commonly target higher turnover on a defined substrate or reaction, using functional assays to rank variants across rounds.
- Reporter-based or growth-based selection for enzyme function
- Quantitative enrichment tracking per variant
- Assay matched to your reaction of interest
Substrate Specificity
Selection can be tuned toward a preferred substrate or away from an unwanted side reaction, depending on what your assay can measure.
- Focused mutagenesis at positions under selection
- Stringency ramped across rounds
- Enrichment data guides the next library
Stability and Robustness
Thermal stability, solubility, and resistance to aggregation are frequently optimized alongside activity for process-relevant conditions.
- Selection under increasing stringency
- Combinatorial recombination of beneficial mutations
- Variants characterized under application-relevant conditions
Where Directed Evolution Fits
Directed evolution is one of several established enzyme engineering strategies, alongside rational design and semi-rational design. Each has specific applications and limitations that should be considered when choosing a strategy for a given industrial process. Directed evolution is typically the entry point when structural information is limited or when the property you need is difficult to predict from structure alone.
The table below summarizes how the main strategies compare on the dimensions that usually drive the choice. Final strategy selection is confirmed during project scoping against your starting sequence, target property, and assay availability.
| Strategy | Starting Information | Typical Use | Considerations |
|---|---|---|---|
| Directed evolution | Functional starting sequence; no structure required | First-round optimization and property discovery | Library size and screening capacity set the pace |
| Semi-rational design | Partial structural or mutational data | Lead optimization around known hotspots | Depends on quality of prior data |
| Rational design | Detailed structural knowledge | Targeted changes at defined positions | Limited by model accuracy for the target property |
| ML-guided loop | Efficient navigation of large sequence space | Requires measured fitness data from prior rounds |
How an Evolution Campaign Runs
A campaign is organized as a sequence of rounds. Each round produces both an enriched population and quantitative data that informs the next round, so the process adapts as it progresses rather than following a fixed script.
Define the Target Property
We agree on the property to optimize, the assay that will measure it, and the selection pressure that will be applied, so that every subsequent round has a clear fitness readout.
Construct the Mutant Library
Diversity is introduced into the enzyme-encoding gene using methods such as error-prone PCR, DNA shuffling, cassette mutagenesis, or focused saturation mutagenesis at selected positions.
Screen or Select Variants
Libraries are screened using the selection system matched to your functional requirement, including plate-based assays, microfluidic or FACS-based sorting, and yeast or mammalian display platforms.
Analyze Enrichment by NGS
Next-generation sequencing resolves which sequences enriched across the round, identifying positions under selection, co-occurring mutations, and where additional diversity is likely to yield further improvement.
Customization Options
Every campaign is scoped against your starting enzyme, target property, and assay capability. The options below describe what can be tailored; the specific combination is defined in the project scope.
Where you already have selection data, we can run a machine-learning-guided loop in which each round is both an experiment and a training set. Where you do not, classical directed evolution remains the fastest route to a real answer.
Mutagenesis Strategy
Random, focused, or combinatorial libraries are matched to your protein and the information available about it.
- Error-prone PCR across the full gene
- Site-saturation and degenerate codon libraries
- DNA shuffling and cassette mutagenesis
Selection Format
The screening or selection system is chosen to match the phenotype your assay can measure.
- Plate-based and microfluidic screening
- FACS-based sorting
- Yeast or mammalian display platforms
Variant Readout
Sequencing depth and analysis scope are set per project to give quantitative enrichment data rather than qualitative colony-picking results.
- NGS-resolved enrichment per variant
- Identification of positions under selection
- Data to guide subsequent rounds
Project Scope Parameters
The table below describes the parameters that are defined case by case when a project is scoped. It is not a fixed package structure; scope, library size, screening depth, and validation are agreed in the project statement of work.
Typical campaigns run two to four rounds of mutagenesis and selection, with improvements often measurable after the first round. The exact number of rounds is decided against your target property and the enrichment data as it accumulates.
| Parameter | Typical project scope | How it is set | Notes |
|---|---|---|---|
| Starting enzyme | Client-provided functional sequence | Reviewed at scoping | A functional starting point is required |
| Target property | Activity, specificity, stability, or combinations | Agreed before library design | Must be measurable by an available assay |
| Library construction | Random, focused, or combinatorial, as scoped | Matched to available structural or mutational data | Method choice follows the protein, not a fixed template |
| Screening format | Plate-based, microfluidic, FACS, or display | Selected against the functional requirement | Display platforms include yeast and mammalian systems |
| Sequencing depth | NGS-resolved variant analysis, scoped per project | Set with the screening format | Provides quantitative enrichment data |
| Number of rounds | Typically two to four rounds | Adjusted as enrichment data accumulates | Stopped when gains flatten or the target is met |
| Validation | Lead variant characterization, as scoped | Defined with the target property | Application testing available where required |
| Deliverables | Engineered enzyme variants with supporting data | Confirmed in the project statement of work | Format and documentation agreed at scoping |
Why Teams Choose This Approach
Directed evolution is an empirical discipline: it does not depend on a perfect structural model, and it produces measured fitness data rather than predictions. That makes it a practical route to enzymes with modified activity, specificity, and stability for industrial processes.
The method is also flexible. It can be run as a standalone campaign or combined with rational and semi-rational design, and it integrates naturally with machine-learning-guided workflows once selection data exists.
Works Without a Model
Directed evolution does not require detailed structural knowledge, which makes it suitable when structural data is limited or the target property is hard to predict.
- Broad mutation followed by selection
- No dependence on docking or structure prediction
- Useful when prior data is sparse
Real Fitness Measurements
Each round returns measured performance on thousands of variants, giving you data rather than predictions to guide the next decision.
- Quantitative enrichment per sequence
- Round-by-round comparison
- Evidence for stopping or continuing
Combines With Design
Directed evolution can be paired with rational or semi-rational design and with machine-learning-guided rounds when training data is available.
- Focused libraries at enriched positions
- Model-guided diversification
- Recombination of beneficial mutations
Engagement and Support
Projects begin with a scoping discussion covering your starting enzyme, the property to optimize, and the assay that will measure it. From there we agree on the campaign design and the deliverables before laboratory work starts.
Support is provided through a named scientific contact at project start, milestone review calls, and email response within one business day. Technical support terms are consistent across projects and are not tiered.
| Item | Detail | Availability | Notes |
|---|---|---|---|
| Technical support | Named scientific contact at project start; milestone review calls; email response within 1 business day | Throughout the project | Consistent across projects |
| Project scoping | Discussion of starting enzyme, target property, and assay | Before laboratory work | Defines the statement of work |
| Progress reporting | Round-by-round enrichment and analysis updates | At agreed milestones | Data shared with the project team |
| Deliverable format | Engineered enzyme variants with supporting data | At project completion | Format confirmed at scoping |
Scientific Background
Enzyme engineering has become an important tool for meeting demand for enzymes with modified activity, specificity, and stability across industrial processes. Reviews of the field describe rational design, directed evolution, and semi-rational design as complementary strategies, each with specific applications and limitations that guide the choice for a given process.
Recent work in biocatalysis has focused on improving enzymes for a broad range of applications, informed by computer modelling and machine learning, and on incorporating new catalytic functionalities. Directed evolution remains a central method in this landscape, particularly where the target property cannot be reliably predicted from structure.
Getting Started
To begin, share your starting enzyme sequence, the property you want to improve, and any assay you already use to measure it. We will review the information and propose a campaign design with defined rounds, screening format, and deliverables.
If you do not yet have a functional starting sequence, that is a prerequisite for directed evolution and should be resolved before a campaign is scoped.
FAQ
What is the difference between directed evolution and rational protein design?
Directed evolution does not require structural knowledge: it pairs random or focused mutagenesis with high-throughput selection to find improved variants. Rational and AI-driven design use structure and data to predict specific mutations. In a modern campaign the two are often run as a loop, where a model proposes where to diversify, selection returns real fitness data, and that data retrains the model.
Do I need to provide a starting enzyme?
Yes. Directed evolution requires a functional starting sequence, because the campaign works by introducing diversity into an existing enzyme-encoding gene and selecting for improved variants. If you do not have a suitable starting point, that should be addressed before a directed evolution campaign is scoped.
How many rounds of evolution are typically needed?
Most campaigns run two to four rounds of mutagenesis and selection. Improvements are often measurable after the first round. The exact number of rounds is decided against your target property and the enrichment data as it accumulates, and we stop when gains flatten or the target is met rather than continuing automatically.
What properties can be optimized through directed evolution?
Common targets include catalytic activity, substrate specificity, thermal stability, solubility, expression level, and resistance to aggregation. In practice, we screen for whatever phenotype your assay can measure, so the achievable targets depend on the availability of a reliable selection or screening readout for the property you care about.
How is variant performance tracked across rounds?
Libraries are coupled to next-generation sequencing so that variants can be tracked quantitatively across selection rounds. This gives enrichment data at sequence resolution rather than qualitative colony-picking results, and it is used to set the next round's selection stringency and to identify positions where additional diversity is likely to help.
Can directed evolution be combined with machine learning?
Yes, when there is selection data to learn from. A model trained on your own selection data proposes where to diversify next, the laboratory round returns real fitness measurements on thousands of variants, and those measurements retrain the model. Where no such data exists yet, classical directed evolution remains the practical starting point.
References
- Victorino da Silva Amatto I, Gonsales da Rosa-Garzon N, Antônio de Oliveira Simões F, et al. Enzyme engineering and its industrial applications. Biotechnology and applied biochemistry. 2022;69(2):389-409. View on PubMed
- Ma E, Chen K, Shi H, et al. Directed evolution expands CRISPR-Cas12a genome-editing capacity. Nucleic acids research. 2025;53(13). View on PubMed
- Silverstein RA, Kim N, Kroell AS, et al. Custom CRISPR-Cas9 PAM variants via scalable engineering and machine learning. Nature. 2025;643(8071):539-550. View on PubMed
- Zarifi N, Asthana P, Doustmohammadi H, et al. Distal mutations enhance catalysis in designed enzymes by facilitating substrate binding and product release. Nature communications. 2025;16(1):8662. View on PubMed
- Bornscheuer UT. Concluding remarks: biocatalysis. Faraday discussions. 2024;252(0):507-515. View on PubMed
- Chen H, Fu W, Yang Y. P450-catalyzed atom transfer radical cyclization. Methods in enzymology. 2023;693:31-49. View on PubMed
Start an Enzyme Evolution Campaign
Share your starting enzyme, the property you want to improve, and your assay, and we will propose a campaign design with defined rounds, screening format, and deliverables.