External laboratory selection checklist

Questions to ask before choosing an external laboratory for a cell-based study

A browser-local buyer checklist for defining one cell-study request, collecting exact laboratory declarations and evidence, checking bounded public records and documenting unresolved selection questions.

SingularCell Research. Reviewed for primary-source scope and unsupported claims. This checklist collects bounded evidence but never qualifies, ranks, recommends or predicts the performance of a laboratory.

How to use this resource

Use this checklist for one request version and one candidate entity and site at a time. Avoid confidential details, preserve missing and conflicting information, and move approved records into your controlled system.

Entries stay in this browser page and are not submitted to SingularCell. Download or storage is not implied.

Define the buyer request first

Draft the stable need that every candidate will answer. This worksheet stays in the browser and does not submit or store what you enter.

Request and intended decision

Record the request identifier and version, research question, intended decision and the people accountable for that decision.

Material and cell context

Identify the exact material, cell model, source or lot, relevant state, sample or matrix and handling constraints.

Measurement and controls

Describe the requested property or measurand, readout, unit, timepoint, method constraints, controls and evidence expectations.

Data and report returns

List required native, processed, metadata, QC, provenance, manifest, report, correction-history and delivery formats.

Operations and timeline

Record sample counts, batches, transfers, permitted subcontracting, start window, milestones, turnaround and dependencies.

Review and risk boundaries

State scientific, quality, security, confidentiality, privacy, IP, legal and commercial reviews and who owns each decision.

Ask one candidate these ten question groups

Each prompt requires a bounded declaration, its exact evidence and visible limitations. Completing the list does not qualify the candidate.

Exact entity, site and rolesAsk which legal entity contracts, which physical site performs every phase, who owns each decision and what changes require notification.Evidence prompt: Link entity, site and responsibility records with versions and dates.
Accreditation and exact scopeAsk for the issuer, program, standard, site, laboratory code, status dates, certificate, complete scope and every proposed activity outside scope.Evidence prompt: Link the issuer-hosted record and map exact scope fields to the request.
Method and measurement contextAsk for the measurand, procedure version, platform, software, matrix, range, uncertainty, controls, references and applicable performance evidence.Evidence prompt: Link method and performance records with site, material, conditions and limits.
Cell resource and sample handlingAsk for cell source and state, authentication plan, receipt checks, storage, mapping, retention, return, disposal and exception ownership.Evidence prompt: Link resource, authentication and handling records without inferring fitness.
Data, metadata and provenanceAsk which native, processed, analysis, QC and report files return, how identifiers map and how transformations, versions and corrections remain visible.Evidence prompt: Link a schema, data dictionary, example manifest and retention boundary.
QC, deviations and correctionsAsk who defines and approves QC, how failures and repeats return, how impact is reviewed and how corrections preserve original records.Evidence prompt: Link procedure versions and a bounded example of deviation and correction history.
Subcontractors and handoffsAsk which external entity and site performs each phase and who sends, receives, reconciles, approves changes and escalates exceptions.Evidence prompt: Link a separate declaration and responsibility map for every external performer.
Security, confidentiality and IPAsk about data locations, processors, access, transfer, retention, deletion, incidents, human-data restrictions, confidentiality and ownership decisions.Evidence prompt: Link scoped declarations for qualified security, privacy and legal review.
Capacity, turnaround and escalationAsk for the time-bound start and capacity declaration, dependencies, reservation state, milestones, reconfirmation trigger and escalation owners.Evidence prompt: Link dated availability, schedule, assumptions and change terms.
Scientific decisions and approvalAsk who defines criteria, reviews differences and uncertainty, approves changes and results, releases records and states the meaning of each signature.Evidence prompt: Link the versioned decision map, reviewer roles and unresolved decisions.

Keep evidence states precise

Use these states to prevent a declaration, link or public record from silently becoming laboratory approval or scientific suitability.

Use these states to prevent a declaration, link or public record from silently becoming laboratory approval or scientific suitability.
State	Exact meaning	Does not establish
CANDIDATE_DECLARED	The candidate supplied a value	Truth, verification or capability
CANDIDATE_EVIDENCE_LINKED	The candidate linked an exact record	Independence, sufficiency or applicability
PUBLIC_RECORD_RESOLVED	The buyer resolved an issuer record at a stated time	Scope applicability or future status
EXACT_FIELDS_ALIGNED	Declared entity, site or scope fields align mechanically	Identity, authority, competence or suitability
REVIEWER_INSPECTED	A named reviewer inspected one exact record	Scientific adequacy or approval
CONFIDENTIAL_NOT_INSPECTED	Evidence is asserted but unavailable to the buyer	Evidence present for the decision
HISTORICAL_OR_EXPIRED	The record does not cover the proposed window	The candidate lacks capability
CONTEXT_DIFFERENT	Evidence covers another site, method, material, range or use	Automatic unsuitability or non-equivalence
CONFLICTING_RECORDS	Candidate and evidence records disagree	Which record is correct
HUMAN_DECISION_RECORDED	A named buyer recorded a scoped selection decision	General approval, guaranteed performance or outcome

Record one candidate evidence summary

Summarize what was actually declared, inspected and decided without publishing sensitive evidence or generating a score, rank or recommendation.

Candidate and exact scope

Record the legal entity, site, request version, declaration version, proposed phases and the period the declaration covers.

Candidate declarations

Summarize the candidate-supplied answers, conditions, exclusions, dependencies and wording that still needs clarification.

Evidence and buyer verification

List exact evidence records, which public sources resolved, retrieval dates, field alignments and records unavailable for inspection.

Differences, gaps and conflicts

Record every different context, missing answer, conflicting record, expired item, unresolved question and assigned follow-up owner.

Qualified review decisions

Record the separate scientific, quality, security, legal and commercial reviewers and each decision's exact scope and rationale.

Selection decision and re-review

Record the accountable buyer's scoped decision, conditions, limitations, effective period and material changes that require re-review.

Precise questions make declarations, evidence, differences and gaps reviewable. They do not qualify, rank or recommend a laboratory or predict whether a study will succeed.

Define the request before comparing laboratories

Create one versioned request that states the research question, intended decision, exact material or cell context, requested measurement, controls, data returns, timeline and review boundaries. Without a stable request, two laboratory answers cannot be compared against the same need.

Keep commercial, scientific, quality, security and legal questions visible as separate review lanes. A lower quote, familiar assay name or completed vendor form must not silently become a scientific-selection decision.

Bind the review to one exact request version
State the intended decision and scientific question
Identify material, cell state and measurement context
Specify required evidence, data returns and limitations
Assign accountable human reviewers before comparison

Separate declarations, evidence and verification

Record what the laboratory says, which exact evidence it links, which public issuer record the buyer independently resolves and what a named reviewer decides as four separate layers. A certificate or accreditation logo is not a substitute for the current scope, site, activity and limitation record.

Bind every record to the exact legal entity, performing site, activity, method and version, material or matrix, range, intended use and date. Preserve historical, confidential, expired, scope-different and conflicting states instead of converting them into a pass or fail.

Name the contracting entity and every performing site
Request the certificate and complete current scope
Resolve public issuer records where practical
Record exact matches, differences and uninspected evidence
Keep the human selection decision separate

Review the whole study handoff

Ask about the exact method, cell resource, sample handling, controls, native and processed data, metadata, deviations, corrections, subcontractors, security boundaries, capacity assumptions and escalation path. Matching names do not establish matching implementations or biological meaning.

Qualified people must decide whether the collected evidence is sufficient for the defined study. The decision remains limited to the versions, site, context and time reviewed and should identify conditions and re-review triggers.

Inspect method, instrument, software and performance context
Inspect cell-resource and sample-handling declarations
Specify raw, processed, metadata and provenance returns
Map subcontractors, transfers, deviations and corrections
Record security, capacity, escalation and approval boundaries

Primary sources

NIST NVLAP — Accreditation vs. Certification ↗National Institute of Standards and Technology · Accreditation is bounded by an exact scope and does not certify products or test data or guarantee laboratory performance.
NATA — Procedures for Accreditation ↗National Association of Testing Authorities, Australia · Within NATA accreditation administration, scope descriptions bind activities, materials, determinations, techniques, methods, limitations and site.
NATA — Testing in support of research and development annex ↗National Association of Testing Authorities, Australia · Within its stated context, the annex addresses agreed objectives, responsibilities, samples, methods, quality control, data, limitations, uncertainty and reporting.
NIH — Authentication of Key Biological and Chemical Resources ↗National Institutes of Health · Key resources such as cell lines can differ between laboratories or over time, supporting exact identity and authentication-plan questions.
NIST — Building Measurement Confidence for Cell Characterization ↗National Institute of Standards and Technology · Cell-measurement confidence depends on the measurement process, variability, controls, references, range, uncertainty and intended application.
OECD — Application of GLP to multi-site studies ↗Organisation for Economic Co-operation and Development · A bounded GLP-context example distinguishing phase and site responsibilities, communications, transfers, deviations, reporting and archiving.
NIST — Research Data Framework, Version 2.0 ↗National Institute of Standards and Technology · Configurable research-data lifecycle, governance, security, responsible-party, transfer, retention, metadata and provenance topics.
W3C — PROV Data Model ↗World Wide Web Consortium · Machine-readable relations among entities, activities and agents, including derivation, revision, attribution, association and delegation.
NIH — Selecting a Data Repository ↗National Institutes of Health · Repository-selection considerations include identifiers, metadata, common formats, provenance, security, integrity, retention and human-data restrictions.

Limitations

This checklist does not determine laboratory quality, competence, qualification, trustworthiness, scientific suitability or likely performance.
Accreditation, certification, certificate, scope or public-directory evidence never becomes a guarantee of study suitability or correct results.
Matching assay, method, instrument, software or cell-line names do not establish matching implementation, biological state or equivalence.
Authentication, controls, QC, validation, provenance, hashes, audit history and signatures remain bounded evidence and do not prove scientific validity.
Security, confidentiality, privacy and IP responses require qualified review and do not constitute an audit, certification or legal opinion.
Capacity and turnaround answers are time-bound declarations and do not prove reservation, feasibility or delivery.
The public worksheet stores nothing and must not receive confidential, proprietary, personal, patient or controlled study information.
Qualified people must review the exact request, declarations, records, differences and limitations before selection and after material change.