Technology that Transforms ‘Uncertainty’ into ‘Certainty’
An innovative protein design platform integrating AI with experimental data
Proteins, which are key molecules for biological function in all organisms are translated from genetic information encoded in DNA. They are generated as linear polypeptides of 20 naturally occurring amino acids, then form three-dimensional structures for biological function. The number of the sequences, for example, of proteins composed of 300 amino acids is 2×10390, which is extremely larger than the number of stars in space, 1023. To explore and find a novel protein sequence in this vast space, we need an unprecedented and phenomenal way.
No more relying on intuition, luck, or time. No more bottlenecks in research and development.
Until now, protein research and development has relied on chance, such as intuition and luck, as well as investments in time and effort. aiProtein technology eliminates this bottleneck in protein development.
aiProtein®：A fully custom-made AI model
RevolKa’s AI
- Bioinformatics
- Protein Language Modeling
- Generative AI
- Low-N Learning
Laboratory Data
Experiments using actual substances and reagents
Observation and validation of real-world behavior
An AI approach is the key to successful protein development
The number of amino acid sequence combinations is enormous, and finding the best candidates using human labor or conventional algorithms takes too much time.
AI uses predictive models to narrow down design possibilities, which can significantly reduce the number and cost of experiments.
Laboratory data can also be used as a teacher. AI calculates it.
aiProtein effectively utilizes laboratory data as training data.
It uses Low-N learning, which allows for predictions even with small amounts of data.
By adding experimental data to AI, we can arrive at the correct answer that was previously unknown.
When designing proteins with diverse characteristics, it is difficult to accurately predict each individual’s behavior (stability, activity, cytotoxicity, etc.) using AI alone.
However, by combining it with laboratory data, we can quickly arrive at the sequence (= correct answer) that expresses the desired protein function.
This is possible because we build custom models rather than a middle-of-the-road model for every protein.
aiProtein is RevolKa’s unique approach, combining AI with actual laboratory data.
Case Study of aiProtein®
30 Years of Research
Solved in 7 Days
Using conventional protein design approaches, it would have taken 30 years to transform a green protein into a bright yellow protein. However, aiProtein achieved this in just 7 days.
（Ref：ACS synth.Biol. (2018) 7, 2014-2022）
A fully custom-made AI model for any protein
aiProtein® is compatible with all molecules composed of amino acids. It has already been successfully used with various formats of antibodies (IgG, scFv, diabody, bispecific antibody, etc.), peptides, enzymes, vaccines, and other biologics, demonstrating the breadth of aiProtein®’s applicability. As we work on modifying a wide variety of proteins, RevolKa has steadily built up its unique know-how and skills, evolving not only protein molecules but also aiProtein® itself every day. We will continue to accelerate this evolution, aiming to expand into unexplored modalities, including gene therapy, in the future.
-
Enzymes
-
Antibodies
-
Various other proteins
Supporting Protein Design with Cutting-Edge AI Technology
-
Amino acid sequences only — “Quick & Easy”
For quick protein design
Antibody design service for lead antibody optimization
- Design up to 3 winning candidate sequences
- Cost-effective
- Fully online / Within 2 weeks
-
From design to validation — “Thorough & Comprehensive”
For one-stop protein development
Synergy of AI and protein engineering — Full-package antibody solution
- One-stop service from design to validation
- AI-powered design prediction
- Fully custom-made
