For Bioprocess Scientists

Smarter experiments.
Fewer runs.

DOE Machine is a response surface methodology platform built for fermentation and bioprocess optimization. Design, analyse, and optimize — all in one place.

See how it works
5
Design Types
12
Max Factors
Auto
Model Selection
PDF
Report Output
Capabilities

Everything the experiment needs

From first design to final report, without switching tools or writing code.

Automated Design Generation
Factorial, CCD, Box-Behnken, Plackett-Burman, and DSD. The right design for your factor count and reactor capacity.
Auto Model Selection
Fits main effects, 2FI, and quadratic models simultaneously. Selects the best using AICc — no manual trial and error.
Multi-Response Optimization
Simultaneously optimize yield, viability, cost, or any combination of responses using composite desirability functions.
Diagnostic Plots
Residual plots, actual vs. predicted, response surfaces, and influence diagnostics — generated automatically with every run.
Uncertainty Quantification
Bootstrap confidence intervals, prediction bands, and probability of meeting specification. Supports QbD and regulatory submissions.
Professional Reports
Full statistical report with executive summary, ANOVA tables, surface plots, and optimal conditions. Ready to share.
Workflow

Five steps from hypothesis to optimum

A structured experimental workflow that eliminates guesswork and maximises the information yield from every run.

01
Configure your design
Enter your factors, ranges, and reactor count. Choose a design type or let the tool recommend one based on your constraints.
02
Download your run sheet
A randomized CSV with exact factor levels for each bioreactor run. Follow the order to minimize systematic bias.
03
Run your experiments
Execute the runs in the lab. Record measured responses directly in the CSV. Missing values are handled automatically.
04
Upload and analyse
Upload the filled CSV. The statistical engine fits models, selects the best, and generates all diagnostic outputs.
05
Read your report
Download the PDF report with optimal factor settings, response surface plots, and confidence intervals. Plan your confirmation runs.
Design Library

The right design for every objective

Six established design families covering screening through full optimization.

factorial
Factorial + Centers
Best for 2–5 factors. Detects interactions. Tests for curvature with center point replicates.
12 runs / 3 factors
ccd
Central Composite
Full quadratic RSM. Gold standard for optimization. Fits response surfaces precisely.
10 runs / 2 factors min · 18 runs / 3 factors
box_behnken
Box-Behnken
Quadratic models without corner points. Ideal when extreme combinations are impractical or unsafe.
15 runs / 3 factors
dsd
Definitive Screening
Efficient 3-level screening for 4–10 factors. Main effects unconfounded with 2FIs.
9 runs / 3 factors min · 13 runs / 5 factors
plackett_burman
Plackett-Burman
Rapid screening for 8+ factors. Maximum information from minimum runs.
8 runs / 7 factors · n runs / n–1 factors
ga
Genetic Algorithm
Evolve optimal media compositions across a large solution space.
3–5 rounds / up to 20 factors
→ Open GA Optimizer

Ready to optimize?

Access is by invitation. Sign in with your account or contact your consultant to request access.

Sign In to DOE Machine