The QC Data You Should Demand From Any Cell Supplier

No vendor will write this page, because every item on it is a question that could disqualify their own datasheet. That is precisely why it is worth having.

The checklist below is what we work through when evaluating a cell product before putting it in front of a client. Most of it is answerable in one email. The value is not in any single item — it is that asking all fifteen tells you within a day whether you are dealing with a supplier who has the data and simply does not publish it, or one who does not have it.

The checklist

#	Item	Why it matters	Acceptable specification	Red flag	Provided as standard?
1	Certificate of analysis for your specific lot	A CoA is lot-specific. A datasheet is a marketing document describing the product in general. They are not the same artifact and are routinely confused	A document naming your lot number, with measured values, test dates and a release signature or equivalent	A “representative” or “typical” CoA, or a datasheet supplied in place of one	Yes, from established suppliers. Ask for it before purchase, not after
2	Cell identity by STR profiling	The single most consequential QC item in cell biology. Misidentified and cross-contaminated lines have corrupted a substantial body of published work	STR profile against a minimum of 13 loci per the ICLAC guidance, matching a reference profile at more than 80 percent, per ANSI/ATCC ASN-0002	No STR data; fewer than 13 loci; no named reference profile; “identity confirmed” with no method	For established lines, usually. For primary and iPSC-derived products, inconsistently. Note STR is for human line authentication and does not apply the same way to every product class
3	Mycoplasma status	Mycoplasma is invisible under routine microscopy, alters cell physiology and gene expression profoundly, and spreads through a facility. It is the most common silent contaminant	Negative by a stated method — PCR is the usual commercial method — tested on the lot you are buying, with the test date	“Routinely tested” without a result for your lot; no method stated; a test date long predating your lot	Yes, near-universally. Check it applies to your lot rather than to the bank
4	Sterility: bacteria and fungi	Distinct from mycoplasma testing and not covered by it	Negative by a stated method with a stated incubation period	Assumed from aseptic technique; no test named	Usually stated on the CoA, often without the method
5	Post-thaw viability with method and timepoint	The number is uninterpretable without both. See post-thaw viability	A percentage with the assay named (trypan blue, AO/PI, flow), the timepoint after thaw, and whether it is a release specification or a typical value	A bare percentage. “High viability.” A trypan blue number presented as equivalent to an AO/PI number	Yes for the number, inconsistently for the method, rarely for the timepoint
6	Purity, against named markers and a named method	“Cortical neurons” without a marker panel is a description, not a specification	Percentage positive for named markers by a named method (ICC, flow cytometry, RT-qPCR), with the timepoint	Purity claimed with no markers; markers named with no percentage; a percentage with no method	Common as a claim, less common with all three elements
7	Karyotype or genomic stability data	iPSC lines acquire culture-adaptive abnormalities with passage. A line with an acquired abnormality behaves differently and can invalidate a comparison against earlier work	A karyotype report, or an array or sequencing-based equivalent, with the passage at which it was performed	No genomic data on an iPSC line; a karyotype from an unstated passage; data on the parent line only when you are buying a derivative	For iPSC lines, commonly. For iPSC-derived terminal products, much less so
8	Passage number at shipment, and expansion headroom	Determines how much use you get from the vial and whether your cells match the passage at which the vendor characterised them	Passage number at freeze, plus the maximum passage at which the vendor’s specifications still hold	No passage stated; “low passage” without a number	Usually stated for expandable products; often absent for terminally differentiated ones
9	Donor information: age, sex, ethnicity, health status	Donor variables are biological variables. They are also the fields most likely to be constrained by your experimental design or your reviewers	The available demographic fields, with a clear statement of what is withheld and why	No donor data at all; donor data for a lineage different from the one you are receiving	Variable. Some suppliers publish it fully, some publish nothing
10	Donor consent scope and use restrictions	Consent terms constrain what you may do and propagate silently to derived products. We have documented at least one commercial lineage carrying a sequencing restriction on ethical-consent grounds that flows through to downstream products	Written statement of the consent scope, explicitly covering whether sequencing and commercial use are permitted	Silence. Restrictions buried in a footnote or a linked document rather than on the product page	Rarely volunteered. Almost always answered when asked directly
11	Licensing class and permitted use	The right to sell a derivative is usually not included in the purchase and is the real gate on any commercial programme. See cell line licensing	A clear statement of research use, fee-for-service rights, and commercial derivative rights, plus who grants each	“Research use only” with no explanation of what a commercial licence would involve or cost; per-SKU restriction lists you have to find yourself	Terms are published. Their consequences for your specific plan are not
12	Functional data appropriate to the cell type	Viability and purity say nothing about whether the cells do their job. For neurons that is electrical activity; for myoblasts, fusion; for cardiomyocytes, beating	Raw or minimally processed functional data with conditions stated: platform, density, coating, co-culture, timepoint. See MEA-ready	A summary bar chart with no n and no conditions; a representative image with no scale bar	No. This is the largest routine gap in the market and the one most worth pushing on
13	Required media, supplements and coating, with the performance context	Datasheet performance is achieved under the vendor’s conditions. If those conditions require their proprietary medium, that is part of the price	A full protocol naming every reagent, concentration and incubation, and a statement of whether the specifications hold on alternatives	Performance figures with no stated culture conditions; a medium described only by catalogue number with no formulation guidance	Yes, as a protocol. The commercial dependency it creates is rarely made explicit
14	Lot-to-lot comparability data	Two lots of the same catalogue number can differ enough to break a longitudinal experiment. This is the item that decides whether your study can span a year	Comparative data across at least two or three recent lots on the specifications that matter to you	No lot-comparison data; no change-notification commitment; an inability to say when the protocol last changed	No. Almost never volunteered. Ask anyway — the answer is informative either way
15	Acceptance window, claim process and remedy	A specification with no consequence is a description. This item converts the other fourteen into something enforceable	A stated window for raising a claim, a stated process, and a stated remedy: replacement, credit, or nothing	No window stated; remedy at the vendor’s discretion; the terms only discoverable after purchase	Terms exist. They are usually short, usually in the general terms rather than on the product page, and worth reading before the shipment lands

The four that matter most

If you only ask four questions, ask these.

Item 2, identity. Cell line misidentification is the field’s most expensive and most persistent quality failure. The standard exists, the reference databases exist, and the test is inexpensive. There is no good reason for a supplier of an established human line to be unable to produce a 13-locus STR profile matching a reference at better than 80 percent.

Item 3, mycoplasma. Cheap to test, catastrophic to miss, and it contaminates your whole facility rather than just your experiment.

Item 12, functional data. This is where the market is weakest and where the gap between suppliers is largest. Viability and purity are commoditised claims that every vendor makes. Raw functional data — a force trace, a spike raster, a fusion index with images — is volunteered by very few, and the ones who volunteer it are telling you something about their confidence.

Item 14, lot-to-lot comparability. The item nobody asks and everybody eventually needs. A supplier who can answer it has a mature process. A supplier who cannot may still have a good product, but you now know to build bridging comparisons into your plan.

How to actually use this

Send items 1 through 15 as a numbered list, before the purchase order, and ask for the answer in writing. It reads as demanding. It is not — every item is something a competent supplier either has on file or can state plainly that they do not have.

Then read the answers for two things:

What is missing, and how they handle its absence. A supplier who says “we do not perform karyotyping on differentiated products, but here is the parent line report at passage 12” is telling you the truth and giving you what they have. A supplier who answers a question you did not ask is avoiding one you did.

Whether the numbers come with conditions. Any specification without a method, a timepoint and a set of conditions attached is not yet a specification. This applies uniformly across viability, purity and function, and it is the difference between a datasheet you can plan against and one you cannot.

On arrival

QC does not end at the purchase order. Four things to do the day the shipment lands, before you commit an experiment to the material:

Check the shipment condition — dry ice remaining, temperature logger if one was included, packaging integrity — and photograph it before unpacking further.
Run your own viability count by the vendor’s stated method, and record it against the CoA.
Seed a small attachment control at the recommended density and count at 24 hours. This is the number that predicts your experiment. If you are not sure what “the recommended density” works out to in your plate format, vials to wells has the published growth areas and the conversion.
If anything is out of specification, raise it within the vendor’s claim window. That window is usually short and it starts on delivery, not on the day you discover the problem.

Keeping the CoA, your own arrival counts, and the shipment photograph together as a single record per lot costs a few minutes and is the difference between a supported claim and an argument.

Where suppliers are genuinely constrained

It is worth being fair about which gaps are refusals and which are structural.

Terminally differentiated products are harder to characterise than lines, because many standard assays assume a proliferating population. Primary cells have inherent donor variability that no amount of process control removes. Functional assays are expensive and platform-specific, so a vendor cannot validate on every buyer’s hardware. Donor demographic data is sometimes withheld for legitimate privacy reasons rather than commercial ones.

None of these are reasons not to ask. They are reasons to accept a well-explained “we do not have that” as a real answer, and to price the resulting uncertainty into your own plan rather than pretending it is not there.

Cell supplier QC checklist: 15 items

The checklist

The four that matter most

How to actually use this

On arrival

Where suppliers are genuinely constrained

Sources