Adaptive Design of Experiments

Get the most from your experiments

From optimising metabolic pathways to designing a high performing reagent mix, let your data guide what experiment to run next to reveal the relationships between variables and make your experiments more powerful.
How it Works

The problem

Navigating the infinite with finite resources is daunting

We know how frustrating it is to spend hours gathering and analysing data only to struggle to find the insight you need. Wrong experiments are the most expensive mistakes one can make.  We’ve been there. We knew there must be a better way.
How it Works

How it works


Adaptive Design of Experiments
gets you faster to your goal
Go beyond brute force high-throughput screening to leverage advanced statistics to navigate the complex design spaces inherent with Biological Systems. Focus only on the most information-rich experiments so you can uncover actionable insights in fewer iterations.
adaptive design of experiments - sampling
01
Design

Identify the most relevant experimental factors and design information rich experiments.

02
Measure

Execute experiments then measure response and variability.

design of experiments - measurement
DoE - modeling
03
Model

Build advanced statistical models that fit and explain your data.

04
Optimise

Leverage mathematical models to design your next set of experiments that maximise response.

DoE - optimisation
01
Design

Identify the most relevant experimental factors and design information rich experiments.

adaptive design of experiments - sampling
02
Measure

Execute experiments then measure response and variability.

DoE- measurement
03
Model

Build advanced statistical models that fit and explain your data.

design of experiments - modelling
04
Optimise

Leverage mathematical models to design your next set of experiments that maximise response.

adaptive design of experiments - optimisation

Focus your experiments on where it's “hot”

Adaptive Design of Experiments is like your data telling you where is “hot” and where is “cold” in your search space.  It balances the exploration of new territories and the exploitation of warm leads between successive rounds of experiments to get you faster to the desired solution, while also exceeding your target.
Our Solutions
Focus on where its hot - DoE

Understand your system,
uncover hidden relationships.

Leverage our cutting-edge machine learning models to build a model that can target and predict a system response in unprecedented detail. By modelling interactions within your system you can uncover insight into the complex relationships between factors and the opportunities hidden within.
Our Solutions

Applications

Pathway Optimisation
Identify the most effective combination of regulatory units that drive improved yield and stability for your target of metabolic pathway engineering.

Growth Medium
Drive improved growth rate all the way to scale up, by understanding the most effective growth conditions for your production organism.

Storage and Transport
Reduce costs and improve long term product performance by understanding the optimal conditions for storage and transport.

Reagent Mix
Improve yield and experimental efficiency by optimising the protocols and formula for your experimental reagent mix.
Our Solutions

Our Solutions

For teams who specialise in the wet lab and want to add cutting-edge biostatistics to their toolbox in fast, simple and cost effective way.
  • You set the goal.
  • Our team identifies the most appropriate advanced statistical method to use. The algorithm generates the first round of experiments for you to run.
  • You run the experiments and share the data with us.
  • The algorithms analyse the data and generate the second round of tests with some exploring the search space and others converging towards promising areas. We repeat this until the results satisfy you - we either do less experiments as we reached the goal or we do the same number of experiments you would have normally done but exceed the target.
For teams with internal biostatistic capacity that want to supercharge their computational systems with effective solutions for discovery, we offer this self-service option.
  • Through our platform, your data analytics team can get access to our algorithms, select the most appropriate to use depending on your need and get them to generate the experiments rounds. We will be available for support, when needed.
  • Once you share your data through the platform, we will make sure it is harmonised and available for use by the algorithms.
For teams who specialise in the wet lab and want to add cutting-edge biostatistics to their toolbox in fast, simple and cost effective way.
  • You set the goal.
  • Our team identifies the most appropriate advanced statistical method to use. The algorithm generates the first round of experiments for you to run.
  • You run the experiments and share the data with us.
  • The algorithms analyse the data and generate the second round of tests with some exploring the search space and others converging towards promising areas. We repeat this until the results satisfy you - we either do less experiments as we reached the goal or we do the same number of experiments you would have normally done but exceed the target.
For teams with internal biostatistic capacity that want to supercharge their computational systems with effective solutions for discovery, we offer this self-service option.
  • Through our platform, your data analytics team can get access to our algorithms, select the most appropriate to use depending on your need and get them to generate the experiments rounds. We will be available for support, when needed.
  • Once you share your data through the platform, we will make sure it is harmonised and available for use by the algorithms.
talk to an advisor

Case Studies

Neural Network models for reagent composition optimization

100X performance improvements
Business challenge

An expert in Cell Free Synthesis faced the challenge to optimize the reagent mix to maximize protein production using cell-free in vitro transcription-translation.

Solution Delivered

Our solution used neural network models to select the most important parameters and identify optimal conditions to boost protein expression.

Impact created

Achieved more than 100X performance over the baseline by exploring only 800 different combinations over hundreds more theoretically possible ones. Additionally, discovered an unanticipated behaviour with a different kinetics that supported longer operations.

Culture medium optimization for protein expression

48% yield improvements
Business challenge

A leader in Protein Synthesis needed to boost protein expression  of a human cytokine without genetically modifying the host organism.

Solution Delivered

OB leveraged its expertise in adaptive design of experiments to explore the  medium formulation space to find alternative media composition to boost protein expression.

Impact created

Improve yield 48% over the baseline while screening only 500 mutants and despite the facts the experimental space was scarce and most of the mutants were inactive.

Discover how you can drive best-in-class performance with Adaptive Design of Experiments