LabKick DOE
Describe the goal and the factors you can vary, and get a justified design and a randomized run sheet. Enter the responses afterwards and get the model, the interactions and the optimum.
Top features
- The design is justified
“Definitive screening design, 17 runs, four continuous factors, catches interactions cheaply without a full 81-run factorial.”
- Constraints in conversation
“Temp 16–37 °C.” “I can only do 24 runs.” “These two cannot both be high.”
- A run sheet you take to the bench
Exact conditions per run, randomized, with center points, flowing into a protocol and the notebook, and into automation if you have it.
- It leads with the answer
“Max yield at 22 °C, 0.4 mM IPTG, OD 0.8; the temp × IPTG interaction is significant, at high temp, more IPTG hurts.”
- What-ifs and the next round
“What if I am stuck at 30 °C?” It predicts the optimum with confidence bounds and proposes runs to confirm or refine.
- The full design catalog
Full and fractional factorial, response-surface, Plackett-Burman, definitive screening, blocking, replication, custom constraints.
Everything in DOE
- Goal stated in plain language
- Factor and range definition
- Constraints between factors
- A run-count budget you set
- Full factorial
- Fractional factorial
- Response-surface / central composite
- Plackett-Burman
- Definitive screening
- Blocking
- Replication
- Center points
- Randomized run order
- Run sheet into protocols and the notebook
- Hand-off to lab automation
- Responses entered, or pulled from Stats
- Model fitting
- Main effects identified
- Interactions identified
- Optimum with confidence bounds
- What-if queries
- Next-round run proposals
- Explicit model terms for people who know DOE
- Bench, cell-culture and bioprocess optimization
What changes
- The two barriers to DOE are choosing the design and interpreting the model. Both become a conversation, and both are AI's strongest suits.
- Most academic labs do one factor at a time and never touch DOE, because the software is expert-only. This is DOE a graduate student actually uses.
- Designing the runs and analyzing them are one tool. In JMP they are two, and the gap between them is where the work leaks.
Run it on your own data.
DOE is not sold on its own. One price gets you all nine tools, and the trial turns every one of them on.
No card to start · One login covers all nine tools · Cancel anytime