Technology
Aureka Biotechnologies is developing a future generation biologic and protein therapeutic discovery platform.

By integrating high-throughput autonomous evolution, single-cell-based functional screening and AI-based drug design and development, each experiment can interrogate millions of therapeutic candidates and generate large-scale, multi-metric data associated with each candidate with respect to function, binding, and developability. The efficient iteration among design, build, test and learn enables us to develop a pipeline that can exponentially accelerate therapeutic design.

By exploring and learning from parallel protein evolution experiments, Aureka's AI platform can decode the language of life and transform therapeutic design from an empirical and trial-and-error approach to truly rational engineering practice.

Single-cell digital biology

High-throughput
functional screening
Big data
AI multi-objective and de novo design
Autonomous
evolution

Decode the language of life

Autonomous evolution
01 High-throughput synthetic biology technologies to mine, synthesize and characterize therapeutic candidates at scale.

We are developing scalable evolution technologies that can mine the protein sequence, structure and function space using libraries with unprecedented high diversity to enable discovery and data generation. Our data-driven protein language model can therefore perform accurate multi-parameter and de novo design at scale.

Single-cell functional screening
02 Single-cell functional screening.

Our proprietary high-resolution and high-throughput single-cell functional screening platform enables us to significantly increase discovery efficiency while generating indexable, content rich data for each cell. This results in abundant, high-quality training sets for our deep reinforcement learning techniques.

GFAARRYK10010100101000001

10100Enantioselectivit

GFAGRTYK10010100101000001

101001ActivitySpecifici

GFAARRYK1001

10100Enantioselectivit

GFAARTWR0

FAGRTYK10010100101000001

0101010Thermostability

GFAGRTYK0100101

101001ActivityS

GFAARTWR0

101001ActivityS

10100Enantioselectivit

0101010Thermostability

FAGRTYK10010100101000001

01001ActivitySpecifici

GFAGGRTYK10010100101000001

10100Enantioselectivit

GFAARRYK10010100101000001

10100Enantioselectivit

GFAGRTYK10010100101000001

101001ActivitySpecifici

GFAARRYK1001

10100Enantioselectivit

GFAARTWR0

FAGRTYK10010100101000001

0101010Thermostability

GFAGRTYK0100101

101001ActivityS

GFAARTWR0

101001ActivityS

10100Enantioselectivit

0101010Thermostability

FAGRTYK10010100101000001

01001ActivitySpecifici

GFAGGRTYK10010100101000001

10100Enantioselectivit

Digitalizing therapeutic discovery
03 Protein geometric language model and deep reinforcement learning.

We employed geometric-aware protein language model and reinforcement learning to drive the design, build, test and learn cycle. With unprecedented data iteration efficiency, our AI solutions allow accurate and rational therapeutic design with desirable attributes.

Coming Soon

The page is under construction...

Return