Contract differentiation providers compared: who will actually hand you the cells

The short answer: the decisive question in this category is not price, protocol or purity. It is whether the provider will hand you physical cells. Several of the strongest organisations here are structurally data companies — their business is running an assay and returning a report, and the cells are an intermediate. That is a legitimate product. It is also a completely different purchase from a box of vials, and it is resolved inside a quote rather than on a website.

The second finding: only one provider in this comparison publishes a timeline, and nobody publishes a price. Curi Bio states 8 to 16 weeks for CRISPR cell line engineering, 4 to 6 weeks for iPSC expansion and banking, and 4 to 8 weeks for directed differentiation. Everybody else quotes privately. That means the productive work you can do before requesting quotes is scope definition, not price research.

The provider comparison

The three questions that set your price

1. Whose line are you starting from?

This is the largest single fork and it changes both cost and legal exposure.

Their line. Fastest and cheapest. The provider already has banked material, an established protocol on that background, and historical QC data to compare your batch against. The cost is that you inherit their donor, their genotype and their licence terms. bit.bio is the clearest case: the differentiation is inseparable from their opti-ox engineered lines, which is exactly why it is fast and consistent, and exactly why genotype choice is not yours.

Your line. Slower and more expensive, because the protocol must be adapted to a background it was not developed on, and because line-to-line differentiation efficiency varies enough that the first attempt is a feasibility study whether or not anyone calls it one. Curi Bio’s “Your Cells, Our Platform” positioning is the most explicit invitation to this route in the comparison.

A third-party line you both source. The common middle path. Budget for material transfer agreement execution — two to six weeks is normal and it parallelises with nothing.

If you are bringing your own line, ask directly whether the quoted price assumes their protocol will transfer unchanged. If the answer is yes, ask what happens if it does not. That scenario is what consumes the budget.

2. What is the physical deliverable?

This question sounds trivial and it is the one that decides the purchase.

Ask explicitly: do I receive cryopreserved vials, live plated cells, a formed 3D tissue, or a PDF?

All four are legitimate. Only three leave you with material you can use next month.

• Ncardia sells plate-ready cardiomyocytes as a catalogue item with a published specification. That is a physical deliverable you can hold a supplier to.
• Cook MyoSite is a cell manufacturer. The deliverable is cells.
• bit.bio, Axol and FUJIFILM CDI sell catalogue vials, so the physical route exists even where custom work does not.
• Curi Bio’s published workflow engineers customer cells into 3D tissues and runs functional assays. The natural deliverable of that workflow is a dataset, and whether they will also hand over the physical tissue is a contract question rather than a listed product.
• eNUVIO states plainly that they do not offer off-the-shelf assays but develop customised ones. Their contract research list includes 3D muscle cultures, neuromuscular junction models and neural spheroids — genuinely hard capabilities — but the business is assay development.

None of this is a criticism. It is a distinction that must be resolved in the first email rather than the last.

3. What counts as success?

A differentiation service without an acceptance criterion is a time-and-materials arrangement with optimistic branding. Get the criterion into the contract as a number and a method.

Ncardia publishes the best template in this comparison — not as a contractual criterion for custom work, but as a product specification you can point at and ask for the equivalent:

• Purity: ≥70 percent cTnT-positive by flow cytometry.
• Identity: cTnT, MLC2v and Actinin-2 by immunocytochemistry.
• Function: MEA response to dofetilide, nifedipine and isoproterenol — three compounds with three different mechanisms, which is a much stronger claim than any single response.
• Safety: mycoplasma.
• Quantity: cell count and viability.

Cook MyoSite publishes the best analytical menu, and it is lineage-appropriate rather than generic: desmin myogenic content, myogenic differentiation by immunohistochemistry and qPCR, sterility to USP <71>, endotoxin to USP <85>, mycoplasma by nucleic acid amplification, bioburden to USP <61>, and flow cytometry. Naming the compendial method rather than the assay category is what makes a criterion enforceable.

Curi Bio’s criterion is functional rather than compositional — tetanic force, fatigue, damage protocols, cardiac contractility and kinetics. For a muscle or cardiac programme where the endpoint is mechanical, that is a stronger acceptance criterion than a marker percentage, and it is the strongest reason to choose them.

Reasonable criteria in rough order of rigour:

1. Marker percentage by flow cytometry. The standard. A marker, a threshold and a named method.
2. Marker percentage plus a negative marker. Purity without residual pluripotency is a stronger claim than purity alone.
3. Functional response. Better still, and the two ends of the spectrum here are Ncardia’s pharmacology panel and Curi Bio’s force measurement.
4. Yield at a stated viability. The commercially decisive one and the most often omitted. “80 percent pure” says nothing about how many cells you get.

Insist on all four where the money justifies it. At minimum insist on 1 and 4, because purity without yield is not something you can plan an experiment around.

Timelines: what the only published numbers imply

Two honest readings.

A full programme from an unedited line to differentiated, QC’d cells is a quarter to two quarters. Editing plus banking plus differentiation, at the published durations, is 16 to 30 weeks before any experiment happens. If your project plan assumes six weeks, the plan is wrong, not the vendor.

These are phase durations, not your durations. They describe an established workflow on cell types that vendor runs routinely. A lineage they do not run regularly, on a background they have not seen, will take longer, and the honest vendors say so at quote stage.

For neuronal work there is a further window that is routinely forgotten: the three-to-seven-week network maturation period after delivery, which sits on top of everything in that table.

Protocol route matters more than vendor choice

Two providers offering “iPSC-derived cortical neurons” may be running fundamentally different processes with different timelines, purities, subtype fidelity and network behaviour.

Transcription factor-driven routes — NGN2 induction being the canonical example, established by Zhang and colleagues as a rapid single-step induction of functional neurons from human pluripotent stem cells — are fast and highly pure. bit.bio’s opti-ox platform sits in this family, and the speed and consistency are the product.

Small-molecule patterning routes such as dual-SMAD inhibition are slower and produce a more developmentally faithful mixed population.

The same fork exists in muscle. Uchimura and colleagues published a MYOD1-based human iPSC myogenic differentiation system explicitly designed to permit high-throughput drug screening — a transcription factor route optimised for consistency and scale — against the slower patterning routes that pass through a PAX7-positive progenitor state.

Neither is better in the abstract. They answer different questions, and the choice determines what your cells will do on an electrode array or in a co-culture. Ask every provider which route they use and why, and be suspicious of any that treats the question as an implementation detail.

Where each provider is genuinely strongest

Curi Bio is strongest where the acceptance criterion is a force number rather than a marker percentage, and it is the only vendor publishing both phase durations and engagement models. The trade is that the natural deliverable is a dataset.

Ncardia is strongest on published product specification. The ≥70 percent cTnT-positive threshold with a three-compound MEA pharmacology panel is the most concrete published QC package in this comparison, and plate-ready cardiomyocytes are a physical product with that specification attached. The trade is that the offering is cardiac-weighted and the broader lineage service is less documented.

Axol Bioscience is strongest where you need a specific disease genotype rather than a specific protocol. A stated portfolio of more than 100 iPSC lines including more than 60 patient-derived lines, with ISO 9001:2015 certified manufacturing and reprogramming from patient biosamples through to made-to-order models, is genotype breadth nobody else here matches. The caveat is real and worth raising early: some donor lineages carry documented use restrictions, and if you plan to sequence, ask before you plan.

bit.bio is strongest where deterministic, fast, consistent differentiation matters more than donor identity, and their published co-culture and MEA protocols on two instrument families mean you can follow a method rather than develop one. The trade-off is structural: the engineering is theirs, which constrains genotype choice absolutely.

Cook MyoSite is strongest when grade matters more than pluripotency. This is a muscle cell manufacturer with over twenty years in muscle-derived cell therapy and phase-appropriate GMP capability, not an iPSC-first shop. If your programme needs GMP-grade myogenic material with a compendial testing package, no other provider here competes.

eNUVIO is strongest on hard, unusual assay configurations. Their published contract research list includes 3D muscle cultures, neuromuscular junction models in both 2D and 3D, and neural spheroids — capabilities that are genuinely scarce, and paired with their own device fabrication. The trade is that they say plainly they develop customised assays rather than selling off-the-shelf ones, so the deliverable is a developed method and its data.

FUJIFILM CDI is strongest where you want a documented, widely cited off-the-shelf lineage rather than a bespoke one, with published network burst and pharmacology data behind it. Two published constraints apply: product restrictions limit the use of some SKUs to provide services to third parties, which is directly relevant if you are a contract research organisation, and pricing is stated valid in the US and Europe only.

When to do it in-house instead

Three cases where contracting out is the wrong call.

You will run it more than a few times with changes each time. Sending a fast-iterating protocol to a contract organisation converts a two-day feedback loop into a three-week one. If differentiation is your research, keep it.

A catalogue product is close enough. Differentiated vials are publicly priced and often cheaper than a bespoke run. iPSC-derived neurons sit in roughly the $333 to $860 per million band depending on vendor and pack size. Price the catalogue before commissioning anything.

You need the cells next month. Nothing in the published timeline table finishes in a month. If the schedule is the constraint, buy catalogue cells and adapt the experiment.

Conversely, contract it out when the failure mode is a lost quarter rather than a lost experiment: GMP-grade production, editing, banking, and anything where documentation is part of the deliverable.

What to put in the RFQ

Send all of this to every provider at once. It is the difference between comparable quotes and a pile of unrelated proposals.

Field 11 is the one buyers skip. Research use is granted almost universally. The right to manufacture and sell something derived from those cells is not, is negotiated separately, and sometimes with a party who is not your service provider. Raise it in the first conversation.

Where to go next

Custom iPSC differentiation services covers the scope questions in more depth. Differentiation protocol routes sets out the transcription factor versus patterning trade-offs. Buy versus differentiate is the arithmetic on whether to commission at all. Cell line licensing for commercial use covers field 11, which is where these programmes most often go wrong late.

How we can help

We put one normalised specification in front of several of these providers at once — same purity criterion, same yield requirement, same deliverable definition, same grade — and push every response onto the same line items. We also ask the two questions buyers forget: what the physical deliverable is, and what the licence permits downstream.

Tell us the target cell type, the cell count, the acceptance criterion you would sign, and whether you are bringing a line. That is enough to open the enquiry.