Ranomics
Rendered ribbon structure of TEM-1 beta-lactamase, a canonical protein engineering enzyme, with highlighted mutation hotspot residues
Service

Custom protein engineering services

Ranomics runs custom protein engineering services end to end: computational design, variant library construction, display screening, and NGS-resolved validation, for biopharma and industrial biotech

Start a project →
DMSYeast DisplayMammalian DisplayNGS
How we choose

Choosing the right protein engineering method

There is no single best method, only the right one for the problem in front of you. The first questions we ask: what are you optimizing, is there data to learn from, and can the property be measured cleanly? The answer points to one of three paths, and they are often combined.

When there is data to learn from

AI design

Some properties already have enough experimental data to train a model. Thermostability is the clearest case: we design candidate sequences computationally, then confirm them empirically. The model proposes, the bench disposes.

When you need the whole landscape

Deep mutational scanning

Scoring every single substitution maps the full fitness landscape in one experiment. For antibodies this is how we run affinity maturation across the CDR loops, and the same map guides where directed evolution should search next.

When function is hard to assay

Directed evolution

When the property cannot be predicted and is hard to measure directly, iterative rounds of mutagenesis and function-based selection get there empirically. It pairs naturally with deep mutational scanning: map the landscape, then climb it.

Core services

Protein engineering approaches for every stage of development

Iterative optimization 01

Directed Evolution

Iterative rounds of mutagenesis and selection to improve protein function. We construct random or focused mutagenesis libraries, screen via display or functional assay, and advance the best performers into the next round.

Activity & stability
Expression yield
Substrate specificity
2-4 selection rounds
Learn more →
Lead optimization 02

Affinity Maturation

Take a confirmed binder and engineer it for tighter binding. Focused mutagenesis libraries around the binding interface, screened under increasingly stringent conditions. Starting from micromolar affinity, we routinely achieve 10- to 100-fold improvements across 2-3 rounds.

10-100x affinity gain
Stringency titration FACS
On-rate / off-rate tuning
NGS + SPR/BLI readout
Learn more →
Variant characterization 03

Deep Mutational Scanning

Systematically measure the effect of every possible single amino acid substitution across your protein. The output is a complete fitness landscape showing how each mutation affects binding, expression, stability, or enzymatic activity. Also applied to drug target biology, mapping how mutations affect drug binding and resistance.

2K-20K variants per run
Fitness heatmap output
Drug target biology
Yeast display + NGS
Learn more →
Computational + experimental 04

Enzyme Optimization

Begins with computational assessment of the enzyme structure to identify positions most likely to improve performance. Structure-guided predictions narrow the mutational search space before wet-lab screening. Computationally prioritized variants are synthesized and screened using custom functional assays.

Structure-guided design
Thermostability & kcat/Km
Industrial & therapeutic
Custom functional assays
Learn more →
What actually decides a campaign

The method is rarely the problem. The assay is.

Most protein engineering campaigns do not fail on the choice of method. They fail on the screen. The situation we see most often is a client who arrives with an assay that looks fine on paper but is not robust enough to survive selection pressure. The moment you push a library through it, the signal collapses, the controls drift, and the enrichment is noise rather than biology.

So before we commit to a campaign, we pressure-test the assay, and we rebuild it if it is not ready. That validation phase is not overhead. It is the difference between a clean fitness landscape and a wasted screen, and it is the single highest-leverage thing a protein engineering partner can do for you. The engineering method is the easy part once the readout is trustworthy.

Workflow

How a protein engineering project runs at Ranomics

Every engagement follows the same disciplined process. Timeline and scope vary by project complexity, but the structure remains consistent.

01

Scoping and assay validation

Technical consultation to define the target, strategy, and success criteria, then we pressure-test the screening assay for signal window, controls, and reproducibility before committing.

02

Library construction

Variant library generated via site-saturation mutagenesis, error-prone PCR, or gene synthesis. Library quality verified by NGS before screening.

03

Screening and selection

Display-based screening (yeast or mammalian) with FACS/MACS selection, or custom functional assay. Multiple rounds for directed evolution campaigns.

04

Analysis and delivery

NGS quantification, enrichment analysis, fitness scoring, and ranked candidate list. Full technical report with methods, data, and recommendations.

How an engagement works

Three sizes of engagement, one underlying pipeline

We initiate every engagement with a scoping call to define your goal. This may include the biochemical or biophysical properties you want to optimize and the success criteria. A bespoke statement of work is generated for your team that matches the goal with the computational and experimental platforms that get your research there. Two fixed-scope programs cover de novo binder discovery. Protein engineering work on a protein you already have, directed evolution, deep mutational scanning, affinity maturation, or enzyme optimization, is scoped as a custom campaign.

Not sure which size fits? Tell us about the protein and the property you want to optimize and we will scope it. All projects may operate under mutual NDA, and target structures, designed sequences, and screening data are never shared or reused.

100%
Binder guarantee on the AI Binder Sprint
Beam TherapeuticsZymergenBluebird BioUniversity of TorontoHarvard Medical SchoolRevolution MedicinesBioMarinCargillMichigan State UniversitySenti BioObsidian TherapeuticsAltimmuneGigaGenVeritasJoyn BioFabric GenomicsBeam TherapeuticsZymergenBluebird BioUniversity of TorontoHarvard Medical SchoolRevolution MedicinesBioMarinCargillMichigan State UniversitySenti BioObsidian TherapeuticsAltimmuneGigaGenVeritasJoyn BioFabric Genomics
FAQ

Protein engineering questions

What protein engineering services does Ranomics offer? +

Deep mutational scanning, directed evolution, affinity maturation, enzyme optimization, variant library construction, and AI-driven design. Variant libraries are built in house and screened on yeast or mammalian display with FACS or MACS selection and NGS readout. Every project is delivered as a ranked variant list with fitness scores, the full NGS enrichment data, sequence files for top candidates, and a technical report.

Can one company take a protein engineering project from design through to validation? +

Yes, and that is how Ranomics is set up. The same team that designs sequences computationally builds the library, runs the display screen, and reads the result by NGS, so there is no handoff between a design vendor and a screening vendor. Scoping and assay validation, library construction, screening and selection, then analysis and delivery all run under one statement of work.

Do you run custom protein engineering programs, or only fixed-scope packages? +

Both. Two fixed-scope programs cover de novo binder discovery: the Binder Pilot is a single-round starter and the AI Binder Sprint is a 6-8 week multi-algorithm campaign. Protein engineering work on a protein you already have, directed evolution, deep mutational scanning, affinity maturation, or enzyme optimization, is scoped as a custom campaign and specified in the statement of work.

Deep mutational scanning vs directed evolution: which do I actually need? +

They are complementary, not competing. Deep mutational scanning gives you the full fitness landscape in one experiment, every substitution scored, which is ideal for understanding a target or maturing an antibody across its CDR loops. Directed evolution climbs toward function you cannot predict, over iterative rounds, and is the right call when the property is hard to assay. In practice we often run DMS to map the landscape, then directed evolution to reach the peak.

Can AI just design my protein instead of screening? +

Sometimes, and we will be honest about when. AI design works when your property has training data and a model that captures it. Protein thermostability is the clearest example, where we design in silico and confirm empirically. It is oversold for de novo catalytic activity, for cold-start problems with no data, and for multi-property tradeoffs where models miss the interactions. When the data is not there, experimental DMS or directed evolution is the faster path to a real answer.

Do I need a validated screening assay before we start? +

Not necessarily, but the assay is what makes or breaks the campaign, so we validate it first. Clients often arrive with a screen that looks fine but is not robust under selection pressure. We pressure-test the signal window, the controls, and reproducibility across rounds, and we rebuild the assay if needed before committing to a full campaign.

What is deep mutational scanning? +

Deep mutational scanning (DMS) is a high-throughput method that systematically measures the effect of thousands of protein variants on function, stability, or expression in a single experiment. Variants are introduced combinatorially, screened via display or functional assay, and quantified by next-generation sequencing. The output is a comprehensive fitness landscape for your protein.

How many variants can you screen in a single DMS experiment? +

A typical DMS experiment at Ranomics covers 2,000 to 20,000 single-site variants across the target protein. Combinatorial libraries for multi-site scanning can exceed 100,000 variants, depending on the region of interest and assay throughput.

Can you engineer enzymes for industrial applications? +

Yes. Ranomics has optimized enzymes for thermostability, catalytic activity, substrate specificity, and expression yield using directed evolution, DMS, and AI-guided design. We work with industrial biotech clients on food-grade enzymes, biocatalysts, and biosynthetic pathway components.

What deliverables do I receive from a protein engineering project? +

Deliverables include a ranked variant list with fitness scores, full NGS enrichment data, sequence files for top candidates, and a technical report summarizing experimental methods, results, and recommendations for follow-up. Raw sequencing data is available upon request.

Start your protein engineering project

Tell us about your protein, the property you want to optimize, and your timeline. We will design a program and send a proposal.

Start a project →