Most iPSC culture troubleshooting fails not because the cause is obscure, but because three variables get changed at once and the culture recovers for reasons nobody can name. This guide runs the diagnosis in order — stabilize the system, classify the failure pattern, isolate the root cause with a controlled test, then lock the fix in place so it does not come back.
Effective iPSC culture troubleshooting is mostly a design problem. The biology in these cultures is not unusually mysterious; the failure modes are well characterized and there are only a handful of them. What makes them hard to fix is that iPSCs respond to almost every variable in the system — density, medium age, coating lot, handling force, incubator position, operator — so any uncontrolled change is plausibly causal and none of them can be ruled out after the fact.
This guide moves from fastest and cheapest checks to slowest and most expensive, changes one variable at a time, and treats “the culture looks better now” as an observation rather than a conclusion. Work it in order and you will usually have a named cause within two passages.
Prerequisite reading: this guide assumes you already know what the system is supposed to do. If you are new to pluripotent culture, or you want the underlying principles rather than the diagnostics, start with iPSC Culture Explained: What It Does and When to Use It, then come back here.
iPSC Culture Troubleshooting Starts Here: The 30-Minute Triage
Run every item below before you change a single reagent. All of them are free or nearly free, several of them resolve the problem outright, and each one you skip becomes a confounder in every experiment that follows. The two stop conditions come first — if either is true, no other troubleshooting is valid until it is resolved.
Stop condition 1 — When did you last test for mycoplasma? If the answer is “I’m not sure” or “more than a month ago”, test now and stop interpreting results until it comes back. Mycoplasma alters proliferation, metabolism and gene expression without visible turbidity, and it is far more common than most labs assume — a survey of NCBI’s RNA-seq archive found mycoplasma sequences in 11% of series analyzed, meaning contaminated data is already circulating in the public record.[1] A culture with an unknown mycoplasma status cannot be troubleshot, only guessed at. For routine screening, a colorimetric one-step kit such as the MycAway™ Plus-Color One-Step Kit (2G) reads by eye in about 30 minutes directly from culture supernatant.
Stop condition 2 — Are you working from a vial with a documented history? Passage number, thaw date, karyotype date, source. If the line has been in continuous culture for an unknown number of passages, or the vial’s provenance is unclear, the most likely explanation for degrading performance is accumulated drift, and every reagent you swap will be a distraction from that. If you need to restart from a line with known provenance, the undifferentiated iPSC lines in the BioHippo cell lines collection are MEF-iPSC (mouse, CD-1, embryo), REF-iPSC (rat, embryo) and SFFV-OCT4 SOX2 MEF iPSC (mouse, embryo) — all rodent, so they suit reprogramming and pluripotency-network work rather than human disease modeling, for which a repository line remains the route.
- Check the medium’s actual age, not its expiry date. How long has the working bottle been at 4 °C, and how many times has it been warmed? FGF2 has a half-life of roughly eight hours at 37 °C and its activity falls substantially over 72 hours under standard culture conditions.[2] Repeated warming of a whole bottle is one of the most common silent causes of progressive differentiation.
- Look at the plate under low power before you look at cells. Are colonies evenly distributed, or concentrated in the center with a bare rim? Edge wells behaving differently from center wells? Distribution problems point to plating technique, incubator handling or evaporation — not to the cells.
- Confirm the incubator, not just its display. Independent thermometer, independent CO2 reading, water pan level, door-opening frequency, and whether your plates sit near the door or on a shelf that others share. Displays drift and are calibrated infrequently.
- Establish whether anything changed at all. New reagent lot, new plate lot, new water source, new operator, new incubator, a service visit, a building event. Ask the whole group, not just yourself. A surprising fraction of “sudden unexplained” culture failures resolve to a lot change nobody logged.
- Watch someone else do the passage, or have them watch you. Aspiration force, pipetting height, time cells spend in suspension, how long plates sit out of the incubator. Handling differences between operators are real, large, and almost never written down.
- Photograph the culture at the same magnification, daily. If you have no baseline images from when the culture was working, you cannot judge whether morphology has genuinely changed or whether you are now noticing what was always there. Start the record today regardless.
iPSC Culture Not Working? Classify Which of Four Failure Patterns You Are In
When iPSC culture is not working, the single most useful piece of information is not what you see but how it is distributed in time. Acute, chronic, intermittent and silent failures have almost non-overlapping cause lists, and classifying correctly eliminates most of the search space before you run a single test.
Pattern A — Acute: it worked last week, it does not work now
A discrete event changed the system. Because the onset has a date, you can bracket it: list everything that changed in the window, and test candidates against that window rather than against the whole protocol. First moves: reagent lot records for medium, coating, dissociation reagent and plates; incubator service log; water system maintenance; recent personnel change. Do not start optimizing the protocol — an acute failure is almost never a protocol design problem, and rewriting the protocol will destroy the boundary that makes this pattern solvable.
Pattern B — Chronic: performance has degraded slowly over months
Gradual decline with no identifiable event usually means the cells themselves have changed. Culture adaptation is progressive rather than sudden, and it is invisible until it crosses a threshold that affects your readout. Differentiation efficiency is often the first thing to go, because it is more sensitive to genomic change than self-renewal is. First moves: thaw an early-passage vial from the bank and run it in parallel against the current culture — the single most informative experiment in chronic decline, diagnostic in one passage. Send both for karyotype and CNV analysis. Treat improved growth as a warning sign, not a success — faster, hardier cells that passage more easily are the classic signature of a culture-adapted variant taking over, exemplified by the recurrent chromosome 20 minimal amplicon that confers a growth advantage in pluripotent cultures.[6]
Pattern C — Intermittent: some plates work, some do not
The hardest pattern, and the one most often misread as bad luck. If the mean is acceptable but the variance is high, something in the system is uncontrolled rather than wrong. Because averaging across replicates hides it, this pattern frequently survives for months before anyone names it. First moves: stop pooling and start stratifying. Plot every well or flask individually against plate position, operator, day, incubator shelf, reagent aliquot and time-since-thaw. Look for structure in the noise before assuming there is none. Edge wells drying faster than center wells, one operator’s plates underperforming, or Monday cultures differing from Thursday cultures are all common and all fixable once visible.
Pattern D — Silent: the culture looks fine but the data do not
Morphology is a coarse readout. A culture can look textbook and still carry a substantial undifferentiated-marker-negative fraction, a partially compromised pluripotency network, or a genomic change with no visual signature. This pattern is dangerous precisely because there is no trigger to investigate — it surfaces as an inexplicable experimental result rather than as a culture problem. First moves: quantify what you have been eyeballing. Run a pluripotency marker panel by flow cytometry rather than by immunostaining, because you need the fraction of positive cells, not a representative image. A culture that is 78% SSEA-4-positive looks identical down the microscope to one that is 97% positive, and will behave completely differently in differentiation.
On the readout specifically: SSEA-4 is a cell-surface glycolipid epitope, not a protein, so it can be stained on live, unfixed cells and read by flow cytometry with no permeabilization — and it cannot be detected by Western blot, since there is no polypeptide to resolve on a gel. Use Anti-SSEA-4 Antibody (MC813-70), which is documented for flow cytometry as well as ICC/IF. The nuclear factors OCT4, SOX2 and NANOG need intracellular flow with fixation and permeabilization if you want the population fraction; immunostaining shows localization, and Western blot gives a bulk average that hides exactly the heterogeneity this pattern is about.
Common iPSC Culture Problems, Grouped by Domain
These are the iPSC culture common problems that account for the large majority of cases, organized by where in the workflow they present. Work the domain that matches your symptom rather than reading straight through.
Attachment and post-passage survival
| Presentation | Candidate causes, most likely first | Test or corrective action |
|---|---|---|
| Few cells attach after passage | Coating incubation too short or plate allowed to dry; coating lot change; plate surface lot change; medium not pre-equilibrated | Re-coat with extended incubation and confirm the surface never dries; run old vs new coating lot side by side on the same day |
| Cells attach then detach in sheets over 24–48 h | Coating degraded or applied unevenly; seeding density too high causing colony contraction; medium change too forceful | Reduce seeding density one step; add medium down the well wall, not onto the monolayer; re-coat fresh plates |
| Widespread death after single-cell dissociation | ROCK inhibitor omitted, degraded or added after seeding[5]; over-digestion; suspension held too long before plating | Confirm ROCK inhibitor is present in the seeding medium from the moment of plating; shorten enzyme exposure; minimize time in suspension |
| Poor recovery from cryopreservation specifically | Slow thaw; DMSO not removed promptly; ROCK inhibitor omitted at thaw; vial frozen at suboptimal density or during a growth trough | Standardize thaw time and dilution; include ROCK inhibitor for the first 24 h; freeze future vials from cultures in log growth |
Morphology and iPSC spontaneous differentiation
| Presentation | Candidate causes, most likely first | Test or corrective action |
|---|---|---|
| Differentiated cells appearing at colony edges | Colonies left past the point where centers become dense; feeding interval too long; local overconfluence | Passage one day earlier; feed daily without exception; remove differentiated regions mechanically before passaging |
| Diffuse differentiation across the whole plate | FGF2 activity loss from repeated medium warming or bottle age; medium stored too long; coating no longer functional | Aliquot medium into single-use volumes and warm only what you need; move to a fresh bottle; re-coat plates |
| Flat, loose colonies with indistinct borders | Seeding too sparse; medium not supporting the state; early culture adaptation | Raise seeding density; verify medium composition and preparation date; run the pluripotency panel quantitatively |
| Colonies pile up into thick opaque centers | Passaged too late; density too high at seeding; feeding insufficient for the biomass present | Shorten the passage interval; drop seeding density; increase medium volume or feed twice daily during peak growth |
Growth rate and metabolic health
| Presentation | Candidate causes, most likely first | Test or corrective action |
|---|---|---|
| Growth slows progressively across passages | Mycoplasma; accumulated genomic change; medium or coating lot drift; chronic mild over-confluence | Mycoplasma test first, then parallel thaw of an early-passage vial; audit lots against the timeline |
| Medium turns yellow well before the next feed | Biomass higher than the feeding schedule assumes; contamination; incubator CO2 out of range | Measure spent-medium glucose and lactate rather than judging by color; verify CO2 independently; increase volume or feed frequency |
| Cells look stressed with no obvious insult | Nutrient depletion between feeds; ammonia accumulation from glutamine breakdown; osmolality drift from evaporation | Assay spent medium for glucose, lactate, glutamine and ammonia across the feed interval to see where depletion actually falls |
| Elevated background cell death in a stable culture | Handling stress; suboptimal density; mitochondrial dysfunction; subclinical contamination | Quantify death rather than estimating it; assess mitochondrial membrane potential and apoptotic markers alongside viability |
Differentiation output
| Presentation | Candidate causes, most likely first | Test or corrective action |
|---|---|---|
| Low yield of the target cell type | Starting culture not uniformly pluripotent; starting density wrong; small-molecule potency or timing off; protocol not adapted to this line | Quantify pluripotency by flow before starting; titrate starting density; verify compound stock age and solvent handling |
| Mixed population with off-target identities | Incomplete exit from pluripotency; patterning window missed; factor concentration too low to be decisive | Stain for residual pluripotency markers and for the intermediate progenitor stage to find where the protocol diverged |
| Protocol works for one line, fails for another | Line-to-line variability in differentiation propensity — expected, not a protocol error | Re-titrate the key steps per line; do not assume published conditions transfer unchanged |
| Yield fell over months with no protocol change | Genomic drift in the source iPSCs; passage-number creep; reagent lot change | Parallel-thaw an early vial and run the differentiation from both; this separates cell drift from reagent drift in a single experiment |
The Variables Nobody Logs
When the obvious candidates are exhausted, the cause is usually something that is real, influential and absent from the protocol document. These are the ones worth checking explicitly.
- Cumulative medium warming. Not bottle age, but how many times that bottle went to 37 °C. Given FGF2’s short functional half-life, a bottle warmed daily for two weeks is not the same reagent as a fresh one. Aliquot on receipt.
- Time out of the incubator. Plates sitting on the bench during a long passage session experience pH shifts as CO2 escapes. A twenty-minute session and a five-minute session are different treatments.
- Plate position and edge effects. Outer wells evaporate faster, changing osmolality and concentrating everything in the medium. In multiwell formats this alone can generate the intermittent pattern.
- Aspiration force and pipette angle. The most operator-variable step in the workflow and the least documented. Shear during medium changes damages colony edges and seeds spontaneous differentiation.
- Time in suspension. The interval between dissociation and plating varies with how many flasks are processed in a session. Cells plated last have been in suspension considerably longer than cells plated first.
- Small-molecule stock handling. DMSO is hygroscopic; repeated freeze-thaw degrades many compounds. A stock that has been opened thirty times is not at nominal concentration.
- Water quality and glassware detergent residue. Rare, but catastrophic and extremely hard to find once other explanations have been eliminated.
- Who did it. Operator is a legitimate variable to test, and testing it is not an accusation. If two people’s plates differ systematically, that is a finding — it localizes the cause to a handling step you can then identify and standardize.
A note on antibiotics. Adding antibiotics to mask a suspected contamination problem is a common response and a poor one. Prophylactic antibiotic use suppresses visible symptoms while allowing subclinical contamination to persist and spread, and can select for resistant organisms.[3] If you suspect contamination, test for it and discard. Antibiotic rescue of a contaminated iPSC line is rarely worth the risk to every other culture in the lab.
Root-Cause Isolation: The Swap Test
Once you have a shortlist, stop reasoning about it and test it. The swap test is deliberately unglamorous: run the suspect condition against the reference condition, in parallel, on the same day, with everything else held constant. It answers the question in one passage and produces evidence you can put in a methods section.
| Suspect | How to swap it | What the result tells you |
|---|---|---|
| Medium lot or age | Same cells, same plate lot, same coating — split across old bottle and freshly prepared medium | If fresh medium rescues, the cause is factor decay or lot variation. Move to single-use aliquots permanently |
| Coating lot or protocol | Same cells and medium — split across plates coated from the current lot and from a reserved reference lot | Isolates matrix quality from everything else. A frequent finding when the material is a tumor-derived extract with high lot-to-lot variation |
| The cells themselves | Parallel thaw of a banked early-passage vial, run alongside the current culture from the same passage point | The most decisive test available. If the early vial performs, no reagent is at fault and you are looking at drift |
| Operator or handling | Two people passage sister flasks from the same suspension on the same day | Localizes the cause to a handling step. Follow with direct observation of the divergent step |
| Incubator | Split sister flasks between two incubators, and between shelf positions within one | Separates equipment fault from position effects. Both are common and neither shows on the display panel |
| Seeding density | Three-point density series from the same suspension, scored at fixed timepoints | Density is line-specific and drifts as growth rate changes. Re-titrating is often the whole fix |
Three rules that make the swap test worth running:
- One variable per test. If you swap medium and coating together and the culture recovers, you have learned almost nothing and will have to repeat this in six months.
- Score quantitatively, at a fixed timepoint. Attachment efficiency, doubling time, percentage of marker-positive cells, differentiation yield. “Looks better” is not a result, and it does not survive contact with a reviewer.
- Keep a reference lot in reserve. A swap test needs something to swap against. Reserving and freezing a known-good aliquot of each critical reagent is what makes future diagnosis possible at all.
iPSC Culture Optimization: Making the Fix Permanent
Finding the cause is half the work. Most labs then fix the immediate problem and change nothing structural, which guarantees a repeat. iPSC culture optimization in the sense that matters here is not chasing marginally better performance — it is reducing the number of uncontrolled variables so that the next problem is diagnosable in days rather than months.
Fix the reagent supply chain
- Aliquot on receipt. Single-use volumes for medium supplements and growth factors, so no bottle is ever warmed twice.
- Reserve and qualify lots. When a critical reagent lot performs, buy through and reserve a reference aliquot. Qualify each new lot against the reference before it enters routine use, not after a failure.
- Log lot numbers in the same place as your results. Lot data that lives only on a discarded box is not data.
Fix the cell supply chain
- Bank deep and early. A large low-passage bank is the single highest-value investment in a pluripotent culture program, and it is what makes the parallel-thaw test possible.
- Define a working passage window for each project, and return to the bank at the ceiling rather than continuing to passage.
- Set a QC cadence and hold to it — mycoplasma monthly, pluripotency panel at defined intervals, karyotype and CNV at bank creation and before data-generating experiments, in line with ISSCR genomic characterization standards.[7]
Fix the measurement
- Track a small number of numbers over time. Attachment efficiency, doubling time, percentage marker-positive, differentiation yield. Plotted longitudinally, these show drift months before it becomes a crisis — which converts the chronic pattern from a discovery into a prediction.
- Include the same reference condition in every run so that between-run comparisons mean something.
- Record passage number, medium, coating, lot and operator alongside every result. Not for bureaucracy — because these are the fields you will need to sort by when something goes wrong.
The variability point, stated once. Differences between donor individuals, genetic stability and ordinary experimental variability all contribute to variation in iPSC-derived models, affecting differentiation potency, cellular heterogeneity, morphology, and transcript and protein abundance.[4] Some of what you are troubleshooting is not a fault to be eliminated but a property of the system to be measured and reported. Knowing which is which is the difference between optimizing and chasing noise.
When to Stop iPSC Culture Troubleshooting
iPSC culture troubleshooting has a cost, and past a certain point the rational move is to stop. Consider the culture unrecoverable and restart from the bank — or reconsider the approach entirely — when any of the following is true.
- A parallel thaw shows the early-passage vial performs and the current culture does not. The line has drifted. No reagent change recovers this. Restart from the bank.
- Karyotype or CNV analysis returns an abnormality. Discard. A genomically abnormal line generates results that are not about your biology, and continuing costs more than restarting.
- Mycoplasma is confirmed. Discard and decontaminate the area. Treating is a last resort reserved for irreplaceable lines, and even then it risks the rest of the lab.
- You have run more than three or four well-designed swap tests without narrowing the cause. At this point the problem is more likely to be in the system’s design than in one component, and a clean restart with a tightened protocol usually costs less than continued diagnosis.
- The pluripotent culture was never the point. If you need differentiated human cells for a downstream assay and iPSC maintenance is simply the obstacle in front of that, the fastest fix is to remove the step. Starting from cryopreserved, characterized iPSC-derived cells eliminates both the maintenance and the differentiation optimization, at the cost of flexibility in cell type and genotype.
That last point deserves emphasis because it is the one labs resist longest. Time spent maintaining a pluripotent culture is only worth spending if the pluripotent state is doing work for you — making a cell type you cannot buy, carrying a genotype you cannot obtain otherwise, or enabling editing. If none of those apply, the troubleshooting effort is being spent on infrastructure rather than on the question.
Diagnostic Reagents for iPSC Culture Troubleshooting
These are catalog items relevant to the diagnostic steps in this guide — assay kits, antibodies, biochemicals and recombinant proteins — grouped by which question they help answer.
| Diagnostic question | Product | Category |
|---|---|---|
| Is the culture contaminated? (routine screen) |
MycAway™ Plus-Color One-Step Mycoplasma Detection Kit (2G) LAMP, colorimetric visual readout, 33 species, 100 CFU/mL, ~30 min, no electrophoresis |
Kits |
| Is the culture contaminated? (maximum sensitivity) |
GMyc-PCR Mycoplasma Detection Kit (2G) Nested PCR, 34 species, 2.5 copies/µL, conventional PCR instrument |
Kits |
| Is the culture contaminated? (documented QC) |
MycAway™ Mycoplasma qPCR Detection Kit (2G) TaqMan qPCR, 183 Mollicutes species, ≤10 CFU/mL, validated to EP 2.6.7, JP G3 and USP 63 |
Kits |
| Is the medium exhausted before the next feed? |
EnzyChrom™ Glucose Assay Kit Spent-medium glucose depletion across the feed interval |
Assay Kits |
| Is metabolic load higher than the schedule assumes? |
EnzyChrom™ Lactate Assay Kit Lactate accumulation as a biomass and stress readout |
Assay Kits |
| Is glutamine limiting? |
EnzyChrom™ Glutamine Assay Kit Substrate depletion in long feed intervals |
Assay Kits |
| Is ammonia accumulating to toxic levels? |
EnzyChrom™ Ammonia Assay Kit Glutamine breakdown product, a common silent stressor |
Assay Kits |
| How much death is actually occurring? |
Cell Counting Kit-8 (CCK-8) Quantitative viability at fixed timepoints |
Assay Kits |
| How much death is actually occurring? |
CellQuanti-Blue™ Cell Viability Assay Kit Non-destructive resazurin readout for time courses |
Assay Kits |
| Are mitochondria compromised? |
JC-1 Mitochondrial membrane potential dye |
Biochemicals |
| Is death apoptotic? |
Cleaved PARP Antibody Downstream apoptosis marker |
Antibodies |
| Is death apoptotic? |
Anti-cleaved Caspase-9 (CASP9) Rabbit Monoclonal Intrinsic apoptotic pathway activation |
Antibodies |
| Is the culture still proliferating? |
Ki-67 Antibody Proliferative fraction |
Antibodies |
| What fraction is genuinely pluripotent? |
Anti-SSEA-4 Antibody (MC813-70) Surface antigen — quantitative by flow |
Antibodies |
| What fraction is genuinely pluripotent? |
Anti-POU5F1/OCT3/OCT4 Polyclonal Antibody Core pluripotency transcription factor |
Antibodies |
| What fraction is genuinely pluripotent? |
Anti-NANOG Polyclonal Antibody Core pluripotency transcription factor |
Antibodies |
| Where did differentiation diverge? |
Anti-PAX6 Antibody Early neuroectoderm checkpoint |
Antibodies |
| Where did differentiation diverge? |
Anti-Nestin Antibody Picoband® Neural progenitor stage |
Antibodies |
| Did neurons actually form? |
Beta III Tubulin Antibody / TUBB3 Pan-neuronal identity marker |
Antibodies |
| Is dissociation survival the problem? |
Y-27632 ROCK inhibitor — single-cell and thaw survival |
Biochemicals |
| Is factor decay the problem? |
Human FGF-2 (154 aa) Protein, His tag (Animal-Free) Fresh supplementation for the swap test |
Proteins & Peptides |
| Is factor decay the problem? |
Recombinant Human TGF-β1 Protein, C-His Fresh supplementation for the swap test |
Proteins & Peptides |
Confirm validated applications on each product page before use — several antibodies above are documented for IHC and Western blot, and immunocytochemistry or flow cytometry compatibility should be verified for your intended readout.
Not stocked — source elsewhere. Several tests named in this guide use reagents BioHippo does not carry: karyotyping, CNV array and STR authentication services, hPSC maintenance media, matrix coatings, dissociation reagents and cryopreservation media. Annexin V apoptosis kits and cell-permeant ROS probes are also not stocked. The genomic characterization checks in particular are among the highest-value diagnostics in this guide, so we would rather name them plainly and point you to a cytogenetics core or commercial testing service than build the guide around what we happen to sell.
Note also that the qPCR kit above uses a separate nucleic acid extraction step; the vendor’s matched magnetic-bead pretreatment kit is not currently stocked, so plan your extraction route before ordering.
Frequently Asked Questions
My culture recovered after I changed several things at once. Do I need to go back and find out which one it was?
If the culture supports work you intend to publish, yes. Not for its own sake, but because an unidentified cause is an uncontrolled variable that will recur — and it will recur at a less convenient moment. The efficient version is a single retrospective swap test against whichever change you consider most likely, run on sister flasks while the culture is healthy. That is far cheaper than diagnosing the same problem again under time pressure.
How do I tell iPSC spontaneous differentiation apart from genuine culture failure?
By distribution and trajectory. Spontaneous differentiation is normal at low levels, appears at colony centers or edges, and is manageable by passaging earlier and removing affected regions. It becomes a failure when the fraction rises passage over passage despite correct handling, when it appears uniformly rather than focally, or when the pluripotent fraction measured by flow cytometry drops below what your downstream protocol requires. Quantify rather than eyeball — that transition is hard to see and easy to measure.
Is it worth testing for mycoplasma if there is no visible turbidity?
Turbidity is not the relevant signal. Mycoplasmas lack a cell wall, pass through standard filtration membranes, resist the antibiotics most commonly used in culture, and reach high concentrations without visible cloudiness. They alter proliferation, metabolism and gene expression while the culture continues to look acceptable.[3] Monthly nucleic-acid-based testing is the standard, and an unknown mycoplasma status invalidates every other troubleshooting conclusion you might draw. Pick the method by purpose: a colorimetric LAMP kit for fast routine screening, nested PCR when you need maximum sensitivity, or a pharmacopoeia-validated qPCR kit for documented, publication- or QC-grade results.
My cells grow faster and passage more easily than they used to. That’s good, isn’t it?
Usually the opposite. Improved growth and hardiness in a long-term pluripotent culture is the classic presentation of culture adaptation — a genetic variant with a growth advantage has expanded and taken over the population.[6] It is welcome in the short term and destructive in the medium term, because those variants frequently differentiate poorly and carry changes that make the line unrepresentative. If growth improves without a deliberate change, karyotype and CNV-test before celebrating.
How many replicates do I need before concluding a change actually helped?
Enough that the effect exceeds the variance you already have, which means you need to know your baseline variance first. Most labs cannot answer this because they have never plotted it. Run the reference condition several times and measure the spread of your chosen readout, then size the comparison against that. This is why the optimization section recommends tracking a few numbers longitudinally — the baseline variance is what makes every future decision interpretable, and it costs almost nothing to accumulate.
Can I troubleshoot a differentiation protocol without fixing the pluripotent culture first?
No, and attempting it wastes the most expensive weeks in the workflow. Differentiation output is highly sensitive to the state of the starting population, so a variable input produces variable output regardless of how carefully the differentiation itself is executed. Establish that your pluripotent culture is stable and quantitatively characterized — pluripotent fraction, growth rate, attachment efficiency — before changing anything downstream. Otherwise you are optimizing one protocol against a moving baseline.
Everything checks out, but my results still vary more than I would like. What now?
At some point you are looking at the system’s intrinsic variability rather than a fault. Line-to-line and batch-to-batch variation are documented properties of iPSC-derived models, not signs of poor technique.[4] The productive response is to stop trying to eliminate it and start accounting for it: include multiple independent differentiation batches, report the spread rather than a single representative result, and size experiments against the measured variance. Chasing residual variance beyond that point consumes time without improving the science.
Next Step
Stuck at a specific step?
Tell us the pattern you are seeing — acute, chronic, intermittent or silent — and what you have already ruled out, and we will point you at the diagnostics that fit. If the pluripotent culture is not the point of your experiment, we will say so and show you the route that skips it.
Talk to a specialistReferences
- Olarerin-George AO, Hogenesch JB. Assessing the prevalence of mycoplasma contamination in cell culture via a survey of NCBI’s RNA-seq archive. Nucleic Acids Res. 2015;43(5):2535–2542. PMID: 25712092
- Song H, et al. Thermostable human basic fibroblast growth factor (TS-bFGF) engineered with a disulfide bond demonstrates superior culture outcomes in human pluripotent stem cells. Biology (Basel). 2023;12(6):888. PMID: 37372172
- Drexler HG, Uphoff CC. Mycoplasma contamination of cell cultures: incidence, sources, effects, detection, elimination, prevention. Cytotechnology. 2002;39(2):75–90. PMID: 19003295
- Volpato V, Webber C. Addressing variability in iPSC-derived models of human disease: guidelines to promote reproducibility. Dis Model Mech. 2020;13(1):dmm042317. PMID: 31953356
- Watanabe K, Ueno M, Kamiya D, et al. A ROCK inhibitor permits survival of dissociated human embryonic stem cells. Nat Biotechnol. 2007;25(6):681–686. PMID: 17529971
- International Stem Cell Initiative; Amps K, Andrews PW, et al. Screening ethnically diverse human embryonic stem cells identifies a chromosome 20 minimal amplicon conferring growth advantage. Nat Biotechnol. 2011;29(12):1132–1144. PMID: 22119741
- International Society for Stem Cell Research. Standards for Human Stem Cell Use in Research — Section 3: Genomic Characterization. isscr.org/basic-research-standards/genomic-characterization
This guide is a general diagnostic framework, not a validated protocol for any specific cell line. iPSC behavior is line-dependent and condition-dependent; confirm all changes against your line’s documentation and your own endpoint before adopting them. Several diagnostics named here — karyotyping, CNV analysis and STR authentication — are not offered by BioHippo and should be sourced from a cytogenetics core or commercial testing service. All products referenced are For Research Use Only (RUO) and are not intended for diagnostic or therapeutic use; verify current specifications, validated applications and intended-use statements on the product page before purchase.
