Sample preparation
Which failures are preventable?
For difficult tissue or demanding assays, where do preparation steps cause repeat work? What does the current method already solve, and how would an alternative be qualified?
Research notebook / September 2026
AI meets biology.
The result still
has to work.
Current priority · Digital biology
We are investigating where engineering can improve the physical work behind a biological result. This is independent research and opportunity validation, not a launched laboratory service.
Three questions guiding the work
Begin with the method.
Listen to the practitioner.
These are starting hypotheses. The research should change when an actual workflow, existing alternative or experiment contradicts an assumption.
Experiment reliability
What makes a result trustworthy?
Which inputs, comparators and acceptance criteria distinguish a useful improvement from an appealing demonstration?
Instrument workflows
Where can engineering remove a burden?
Which repeated test, handling or measurement step consumes operator time? What equipment, interfaces and expert judgment does it depend on?
Two tests, both necessary
A result that works.
A problem worth solving.
Scientific and commercial evidence answer different questions.
The method needs representative physical inputs and an accepted comparison. The opportunity needs a responsible buyer, a costly current problem and a reason to choose an improvement over existing options.
Research is in progress. We are seeking practitioners to challenge these questions and help define useful tests.
Selected reading
Sources behind
the questions.
Public work informing our research. The organizations below are not Plixo partners or customers.
Isomorphic Labs · Max Jaderberg · August 2023
Modelling the invisible world of molecular biology with machine learning
Describes bringing machine learning experience into biology through scientific immersion and interdisciplinary work. Our takeaway: learn the domain with practitioners before choosing the technical solution.
Opentrons · Existing vendor services
Custom protocol development and its boundaries
Opentrons offers custom protocol development, code validation and setup guidance for its own hardware. Third-party equipment is outside that service's scope. Our takeaway: compare existing vendor support first, then establish whether a remaining workflow or integration gap justifies new work. Code validation and biological performance need distinct acceptance criteria.
Isomorphic Labs · IsoDDE · February 2026
Drug Design Engine technical report
Evaluates molecular predictions against defined datasets and experimental reference data. Our takeaway: name the comparator and test conditions. Benchmark performance does not establish a Plixo capability.
The next step
What does the work
look like in your lab?
A recent example, a current workaround,
or a reason our
assumptions are wrong.