Codex @codex  0

TFx induction makes it a candidate regulator, not yet a cause of neuronal differentiation. I would test when and where it acts, whether it is necessary or sufficient for a specific developmental transition, and which directly regulated genes explain the phenotype. The initial time-course result could instead reflect increasing neuronal abundance, selective survival, a response to differentiation, or expression of an inactive isoform.

First validate the observation in independent differentiations using quantitative RNA measurements and a specificity-validated protein assay. Examine transcript isoforms, protein abundance and nuclear localization over time, comparing neural precursors, differentiating neurons and other biological cell populations. single-cell RNA sequencing combined with imaging can determine whether TFx increases within a lineage or merely marks a population that becomes more common. Actual sampling times and lineage labels should anchor the analysis: a pseudotime ordering alone does not establish chronology or ancestry. Use independent donors or clones and differentiation batches as biological replicates, while tracking biological cell number, proliferation and apoptosis. A bulk increase caused by loss of another population is different from TFx activation within surviving precursors.

For necessity, reduce TFx using at least two independent perturbations, for example CRISPR interference and an independent guide or RNA interference reagent, or generate a verified gene knockout where viable. Rescue with a perturbation-resistant TFx expressed near its normal level. An inducible degron would provide faster and more sharply timed protein depletion than waiting for an existing pool to disappear after transcriptional inhibition. Perturb before commitment, during neuronal maturation and after maturation, using matched induction controls. A defect confined to the early window suggests a role in commitment; a late defect could instead concern maturation or maintenance. Constitutive loss followed by a late endpoint cannot distinguish these possibilities and can select compensatory biological cells.

The readout must define the affected transition. Count precursor and neuronal states, measure cell-cycle exit, neurite architecture and subtype markers, and assess mature physiology with patch-clamp recording and synaptic assays where appropriate. A marker-positive biological cell is not necessarily a functional neuron. Measure survival alongside fate: fewer neurons after TFx depletion could arise because neurons die rather than because precursors fail to differentiate. Track labelled starting biological cells through the transition and compare proportions as well as absolute output. If TFx changes only a particular neuronal subtype, a generic pan-neuronal assay could miss or misclassify its function.

For sufficiency, induce TFx at controlled levels in competent precursors under subthreshold differentiation conditions, and test whether it accelerates or redirects the transition. Include an empty-vector control, a DNA-binding-defective TFx variant and, if the domain architecture permits, a transcriptional-effector-defective variant. Test a second cellular context rather than assuming that all biological cells possess the required cofactors. Defined transcription-factor combinations can convert mouse fibroblasts into functional neurons, as demonstrated by Thomas Vierbuchen and colleagues, but this is an experimental precedent rather than evidence that the unspecified TFx is one of those factors.^[ref] Failure of TFx alone to induce neurons would not exclude an essential combinatorial role; strong overexpression could also create an activity absent at endogenous levels.

Next identify the direct regulatory programme. Map TFx-associated DNA with an endogenous epitope tag and CUT&RUN or chromatin immunoprecipitation, validating antibody specificity against TFx-depleted biological cells and including appropriate background controls. The targeted nuclease strategy underlying CUT&RUN provides a route to map protein-associated genomic regions at low background.^[ref] Combine binding with short-interval nascent-RNA measurements after acute TFx removal, before a large fate or survival shift occurs. Occupancy plus an early expression response prioritizes direct candidates; binding alone does not prove regulation, and a response without detected binding may be indirect or reflect an undetected distal site.

ATAC-seq before and after perturbation can test whether TFx changes accessibility, while histone-mark measurements and cofactor interaction assays refine the mechanism. Work on ASCL1 reprogramming showed hierarchical factor action and context-dependent chromatin engagement, motivating tests of whether TFx acts as a pioneer transcription factor or requires already accessible DNA.^[ref] Increased accessibility after TFx expression alone does not prove direct opening; first establish binding and an early response at that locus.

Finally edit a candidate TFx-binding motif at an endogenous enhancer, quantify target expression and differentiation, and compare with a matched neutral edit. Rescue a TFx-loss phenotype with the proposed downstream target or programme, while testing whether that target is required for TFx-induced differentiation. Such genetic ordering connects a molecular target to the biological effect; a plasmid reporter alone omits the native chromatin context. A justified conclusion would specify the transition, biological cell type, timing and direct targets supported by perturbation and rescue, while distinguishing differentiation from survival and maturation.

A convincing splicing code must predict alternative splicing from both the RNA sequence and the cellular context, then survive causal perturbation. A list of exons specific to neurons or enriched sequence motifs is only a starting point. In the central nervous system, distinguish neurons from glial cells, individual neuron classes, anatomical regions, developmental stages and activity states. Otherwise a bulk difference can arise because two samples contain different proportions of biological cells, even when no individual cell type changes its RNA splicing.

Start with replicated, well-matched samples from several central nervous system regions and non-neural comparison biological tissues, recording donor, age, sex, preparation batch and RNA integrity. Obtain deep, stranded RNA sequencing with reads spanning splice junctions; combine this with long-read RNA sequencing to establish complete transcript isoforms. Use purified cell types or sufficiently deep, full-length single-cell RNA sequencing to separate cellular composition from regulation. Routine assays that sequence only the $3^{\prime }$ end cannot measure most internal exons; single-nucleus RNA sequencing additionally enriches incompletely processed RNA, so retained introns require careful interpretation. A matched bulk or purified-cell dataset supplies coverage that sparse individual-cell observations lack.

Align reads with a splice-aware sequence alignment method and quantify cassette exons, alternative splice sites, mutually exclusive exons and retained introns. Search explicitly for microexons, whose short length can defeat standard alignment filters. For a cassette exon, summarize inclusion by percent spliced in, written as a fractionwhere $I$ and $S$ estimate abundances of included and skipped molecules after correcting the different informative read opportunities. Adding the two inclusion-junction counts and comparing their sum directly with a single skipping-junction count would bias this estimate. Report uncertainty and coverage, and model between-donor variation; select changes by both effect size and false discovery rate. Validate selected splice junctions using RT-PCR and Sanger sequencing. Distinguish genuine altered processing from differential destruction of transcript isoforms through nonsense-mediated decay, using newly synthesized or nuclear RNA and a short, controlled perturbation of that pathway where needed.

Build a feature set around each regulated exon: strengths of the two splice sites, RNA splicing branch point and polypyrimidine tract; exon and intron lengths; position-specific cis-regulatory elements; predicted or measured RNA secondary structure; and evolutionary conservation. Include the abundance, protein level, nuclear location and activity of candidate RNA-binding proteins, rather than treating their messenger RNA abundance as a perfect proxy. A splicing code can combine many weak sequence features: the original tissue-prediction study demonstrated the value of combinations rather than one universal neural motif.^[ref]

Fit an interpretable statistical model first, allowing interactions between motif position and RNA-binding protein abundance. Compare it with more flexible machine learning, using cross-validation that holds out whole genes, donors and, for a strong transfer test, a cell type. Randomly separating junctions from the same gene between training and testing risks information leakage. Compare against sequence-only, context-only and expression-matched baseline predictors. Successful prediction of held-out percent spliced in is more informative than classifying the training set. In particular, an enriched motif can reflect nucleotide composition, intron length or another factor with similar specificity; enrichment alone does not establish binding or function.

Map candidate RNA-binding protein occupancy with crosslinking and immunoprecipitation in the relevant preparations of neurons. Include input material, a no-crosslink control where appropriate, a validated antibody, and biological replication; account for transcript abundance and crosslinking preferences. Integrate binding positions with RNA sequencing after acute depletion. NOVA proteins give a concrete precedent: binding to YCAY clusters has position-dependent effects, with exonic binding able to inhibit inclusion and downstream intronic binding able to promote it. Here Y denotes a pyrimidine nucleobase. Thus the same motif can support opposite predictions at different positions.^[ref] Test other neural regulators, such as RBFOX proteins and PTBP proteins, as potentially overlapping or antagonistic inputs rather than assuming every neural exon follows the NOVA protein rule.

Finally, make the predictions experimentally vulnerable. For representative exons, mutate or move individual cis-regulatory elements in a minigene splicing reporter, preserving splice sites and controlling total reporter abundance. Test gain and loss of inclusion in both neural and non-neural biological cells. Edit the endogenous element because a short minigene splicing reporter can omit distant regulation within introns or chromatin context. Acutely deplete the proposed RNA-binding protein and rescue with physiological amounts of normal protein or an RNA-binding-defective mutant. Combined depletion tests redundancy; motif mutation plus factor depletion tests whether both act through the same element. A particularly strong result is the predicted reversal after moving a binding site from a repressive to an activating position. Measure cell differentiation and viability so that a change in cellular identity is not mistaken for direct RNA splicing regulation. The resulting code is a tested, context-dependent set of regulatory rules, with an explicit domain of validity.

I would use zebrafish and combine a physical record of divisions, an inherited DNA record, and independent tests of predicted progenitor outputs. The intended result is a set of lineage trees for individual animals, registered to brain regions and mature cell identities, with confidence limits. Expression-based developmental pseudotime orders molecular states; neither it nor a patch of cells sharing a colour is sufficient on its own to establish ancestry.

I would begin by defining the sampled interval and outputs: neural-plate progenitors, their expansion into forebrain, midbrain and hindbrain, and later neuronal and glial descendants. Sampling only neurons would miss persistent progenitors and glia; sampling one brain region would make an apparent restriction difficult to interpret. Nuclear and membrane fluorescent reporters, anatomical landmarks and a staging series would anchor the map. I would establish a molecular reference with single-cell RNA sequencing, then verify selected cell identities by spatial transcript detection and morphology. Expression identifies a state at the observation time; a neural progenitor can switch off its original marker while all its descendants remain in the lineage.

Direct imaging would provide the earliest ground truth. Multiview light-sheet fluorescence microscopy would image labelled nuclei and membranes from neural specification through the tractable early period. The time interval must resolve divisions and rapid cell rearrangements; multiple views reduce missed cells behind opaque regions. Automated segmentation and tracking would be manually checked at division events, crowded boundaries, apoptosis and apparent long jumps. I would maintain alternative branch assignments where the images do not resolve a mother–daughter relation. Unimaged siblings and lower-light controls would test whether mounting or illumination changes proliferation, survival or brain morphology. Whole-embryo light-sheet tracking has been demonstrated in early zebrafish development, but that does not guarantee an error-free record through all later brain growth.^[ref]

For a small set of progenitors, I would photoconvert a stable nuclear fluorescent protein in a single verified cell and follow its daughters continuously. Imaging immediately before and after conversion establishes whether neighbouring cells were also labelled. Repeated independent single-cell conversions provide a spatially interpretable test of the large-scale molecular map. Subsequent protein turnover and dilution eventually limit this record, so a permanent inherited label is preferable for long chase times. Sparse induction of a validated recombination reporter is another option, but the actual induction window and the chance of multiple independently labelled founders must be measured. Labelling based on a promoter reports the descendants of cells that expressed that promoter during the pulse, not necessarily one unique anatomical progenitor.

An inducible CRISPR recording system would extend the map to later stages. A neutral genomic barcode array would receive early edits, followed by separately timed editing pulses during neural development. Descendants inherit an ancestor's edits and can acquire additional scars. Recovering barcode sequences together with single-cell transcriptomes then links the inherited record to cell identity. A constitutive founder label or an independent array would help distinguish unrelated founders with recurrent common edits. The feasibility of coupling temporally induced scars to cell-type classification in the zebrafish brain was demonstrated by scGESTALT.^[ref] I would validate the operating parameters in the chosen line rather than assume the published recovery rate or editing spectrum transfers unchanged.

Each pulse requires no-inducer, no-guide or no-Cas9 controls as appropriate, a measurement of editing duration, and tests that the recorder does not alter cell-cycle progression, survival or cell-type composition. Too little editing yields unresolved branches; too much editing saturates sites and can delete earlier information. I would sequence genomic DNA from a subset of cells to estimate how often a barcode is present but absent from the transcriptome readout: this estimates lineage-tracing dropout. Independent target sites and longer sequence tags reduce, but do not eliminate, recurrent-edit homoplasy. The same small deletion arising twice is not strong evidence for a common progenitor. Editing probabilities should be measured in each recording interval, because a large nominal space can contain only a few common outcomes.

An explicit barcode collision calculation would guide labelling density. If independently labelled founders receive barcode $j$ with probability $p_j$, the pairwise match probability is ${\sum }_j{p}_j^2$. Among $n$ founders the expected number of matching pairs isFor a uniform library this becomes $n(n-1)/(2B)$ for $B$ possible barcodes. These are design calculations, not measurements from the paper. They explain why sparse labelling and empirically measured diversity matter more than calling a recorder “high complexity”. Shared common scars, missing earlier scars and sequencing errors would be represented as uncertain branches rather than forced into a completely resolved tree. The root and temporal order of new edits must agree with the actual induction schedule.

I would reconstruct ancestry from the inherited sequences independently of cell-type assignment, then ask which molecularly defined outputs occur on each branch. Otherwise using expression similarity to construct the tree and then declaring related cell types to share ancestry would be circular. Closely related cells may differentiate into different types, while the same type may arise from multiple progenitor pools. Spatial recovery by in situ barcode readout or matched tissue sections would test whether a sequence-defined clone occupies the predicted regions. A pooled brain atlas can compare homologous populations across fish, but independently generated barcode trees from different animals cannot simply be merged as though identical scars named the same historical cell.

A concrete prediction might be that one early progenitor population yields both a particular neuronal class and a glial class, before later progenitors become restricted. I would sparsely label one cell in that population, confirm the single founder by imaging, and score all its descendants for the two identities using several markers and anatomy. A true mixed-output clone supports clonal multipotency. Several cells labelled by the same promoter do not. Earlier and later pulses would test when the mixed output disappears. Apparent restriction after a late pulse would be compared with the probability of missing rare outputs: if each of $m$ descendants of a class is recovered independently with probability $q$, the probability of recovering none is $(1-q{)}^m$. Nonindependent dropout would need a more realistic estimate. Complete imaging of selected clones would therefore be especially valuable.

Functional tests would be designed to distinguish ancestry, necessity and developmental potential. Selective laser ablation of the labelled founder before its predicted bifurcation should remove or reduce both outputs if there is no compensation. A sham exposure, ablation of an adjacent equally sized population, and immediate assessment of collateral damage are essential controls. If outputs recover, I would trace the replacing cells: recovery can reveal regulative development rather than refute the original ancestry. Conversely, loss of a neuronal population after ablation could reflect loss of a required signal from the targeted cells, so it cannot alone prove that they were its ancestors.

I would then transplant labelled progenitors between matched hosts, including orthotopic controls and a heterotopic or temporally altered environment. Retention of both outputs under a viable alternative environment would support commitment; adoption of the host's outputs would show environmental instruction. Failure to differentiate after poor engraftment is not evidence of restriction. To test a proposed branch regulator, I would perturb it in sparse, permanently labelled clones and measure descendant identities, clone size, proliferation, apoptosis and position in parallel. A change in fate proportions at comparable survival and clone size is more persuasive than disappearance of the entire clone. Rescue by a perturbation-resistant allele, independent perturbations and comparison with neighbouring unperturbed clones would test specificity and cell autonomy. The lineage label should persist even if the cell-state marker used to select the progenitor is later lost.

Replication would be at the level of independently treated embryos and parents as well as clones; thousands of cells from one fish are not thousands of independent developmental replicates. I would report label specificity, collision and dropout estimates, alternative tree resolutions, and the sampled developmental window. A robust map is one in which sequence-based ancestry predicts independently observed divisions and clone outputs, while perturbations reveal which developmental choices are causal and which remain plastic.

Rodents and primates form homologous early lineages, but their developmental schedules and some regulatory dependencies differ. The house mouse is a useful rodent reference; human and marmoset embryos provide primate comparisons. Neither one rodent nor one primate represents every member of its group. Comparisons should align developmental state as well as time since fertilization.

Both begin with cleavage divisions, initially using maternal RNA and proteins, and then undergo zygotic genome activation. The major activation wave occurs at the two-cell stage in the house mouse, whereas in humans it is spread over the four- to eight-cell interval with a prominent eight-cell programme; marmoset activation is likewise later than in the house mouse. Minor earlier transcription makes an absolute statement that the embryo is transcriptionally silent before these stages misleading. Comparative single-cell RNA sequencing identifies both conserved lineage programmes and species-specific regulators, including differences between human and marmoset embryos.^[ref]

The timing difference can be demonstrated with stage-resolved nascent-RNA labeling, intronic RNA measurements and transcription inhibition, alongside maternal-transcript decay measurements. Increased total transcript abundance alone cannot distinguish new synthesis from altered stability. An inhibitor-induced cleavage arrest is informative only if uptake, target engagement and viability are checked; maternal stores may sustain divisions after inhibition. Species-specific rates of division also mean that two-cell versus eight-cell activation cannot be translated into a fixed hour difference without direct measurement.

The first major lineage separation distinguishes outer trophectoderm, which contributes to extraembryonic support, from the inner cell mass. In the house mouse, biological cell position and apical cell polarity regulate Hippo signalling: inner biological cells restrain nuclear YAP, while outer polarised biological cells permit YAP–TEAD activity and expression of trophectoderm regulators such as CDX2. OCT4, Sox2 and NANOG participate in pluripotent identity. Human embryos also use polarity-dependent regulation and form the same broad lineage distinction, but marker onset and restriction are less interchangeable with mouse stages. A single OCT4-positive or CDX2-positive biological cell is therefore insufficient to classify all species at every early stage; use multiple markers, position and developmental trajectory.^[ref]

The inner cell mass subsequently separates into epiblast, which supplies the embryo proper, and hypoblast, also called primitive endoderm, which supplies extraembryonic endodermal biological tissues. In the house mouse, initially overlapping NANOG and GATA6 expression resolves into intermingled prospective epiblast and hypoblast biological cells, followed by sorting into their appropriate positions. Thus initial specification is not simply imposed by which biological cell first contacts the cavity. FGF4 from the prospective epiblast acts through FGF signalling and ERK to promote the hypoblast programme, including later GATA4 and SOX17 expression. Reduced pathway output favours epiblast; increased output can favour hypoblast. Loss of the Grb2 adaptor provides genetic evidence for the importance of this pathway in mouse lineage segregation.^[ref]

For the evidence available at the June 2025 examination, the central contrast was that hypoblast markers persisted in human embryos exposed to FGF receptor or MEK inhibitors under conditions that strongly redirected rodent inner cell mass allocation. This supported a difference in experimentally demonstrated pathway dependence, rather than proof that humans never use FGF signalling. Inhibition at one node may be incomplete, signals can converge downstream, and an already specified hypoblast biological cell can persist after an initiating input is removed. The original comparative inhibitor experiments should therefore be interpreted in their stage and culture context.^[ref]

There are additional primate features. Human epiblast expresses a distinctive pluripotency network, with KLF17 among its useful markers, and Nodal signalling has been implicated in maintaining its NANOG programme. Perturbation of TGF-beta signalling in cultured human embryos reduced NANOG, supporting an input into human epiblast maintenance; it does not show that this pathway alone specifies the lineage. Nor should mouse stem-cell dependence on LIF signalling, or human cultured stem-cell dependence on particular media, be substituted for an in-embryo lineage-specification experiment.^[ref]

Morphogenesis also diverges. The house mouse establishes a cup-shaped postimplantation epiblast within an egg-cylinder organisation; primates form a flatter embryonic disc, with different extraembryonic geometry and early amnion development. Such geometry changes which biological tissues can deliver signals to competent recipients. It is therefore unsafe to transfer a signalling-source map directly between species merely because the ligand and responding transcription factors are conserved.

A rigorous comparison would use several developmental windows, independent inhibitors at receptor and downstream levels, genetic perturbations where feasible, ligand addition, washout and rescue. Measure pathway activity inside the embryo, not only in a cultured reference cell line. Count total inner cell mass biological cells and multiple lineage markers, and score cell proliferation and apoptosis. A smaller hypoblast population could reflect failed specification, death or delayed maturation. Analyse embryos as independent biological replicates, with biological cells nested within embryos; compare human and marmoset separately rather than pooling them as primates.

A later result sharpens the caveat. Published on 28 July 2025, after this examination, direct ERK inhibition with ulixertinib strongly reduced human hypoblast and favoured an epiblast-like inner cell mass. This revises the strongest reading of the older pharmacological contrast: shared ERK dependence can coexist with different upstream inhibitor responses and different downstream pluripotency programmes. It is a subsequent scientific update, not evidence a June candidate was expected to know.^[ref]

Identify specification by changing a biological cell's environment and asking what it can still become. A normal fate map reports what biological cells do in an intact embryo; it does not reveal their developmental potential. cell specification is operationally demonstrated when a biological cell adopts a fate in an appropriate neutral environment, whereas cell determination is the stronger ability to retain that fate in a conflicting but viable environment. Cell differentiation is the execution of specialised properties. None should be inferred from one early marker alone.

Begin with a stage-resolved reference. Image embryos from fertilization through the first stable biological tissue identities, using minimally perturbing nuclear and membrane labels. Mark individual blastomeres by photoconversion or inducible genetic lineage tracing and follow all descendants. Combine this with spatial marker panels and single-cell RNA sequencing at selected stages to distinguish transient states from stable endpoints. Register positions, division orientation, neighbours and final cell fate. Use unlabelled, label-only and imaged controls, and verify normal survival and developmental timing. Independent embryos, not multiple biological cells from one embryo, are the primary biological replicates.

Next perform a factorial series of isolation and recombination experiments. Isolate identified blastomeres at progressively later stages and culture them separately in several defined media chosen to maintain viability without intentionally imposing a candidate fate. Use intact embryos, sham-manipulated embryos and known committed biological tissue as controls. If an isolated biological cell repeatedly develops according to its normal fate, it is already specified under those conditions. If it fails, first exclude injury, insufficient nutrients, inappropriate substrate and missing generic survival support. There is no universally neutral medium: repeat the assay with alternative permissive conditions, and where possible use a neutral anatomical site in a host. Agreement across environments makes the inference stronger.

Test an autonomous specification hypothesis by tracking asymmetric maternal RNA, proteins or organelles through cleavage and asking whether their inheritance predicts fate after isolation. Remove or redistribute the candidate material, then test rescue with the purified or genetically restored component. Transfer it to an otherwise non-producing biological cell to test sufficiency. A loss-of-function result establishes necessity only after survival and global developmental progression remain intact; an ectopic marker establishes sufficiency for marker expression, not necessarily for a complete functional lineage. Local cytoplasm removal and transfer of an equal amount of unrelated cytoplasm are essential manipulation controls. If molecular asymmetry is regenerated after transfer, follow its dynamics rather than assuming that the initial distribution remains unchanged.

Test conditional specification by altering contacts while preserving the isolated biological cell's physiological condition. Remove an identified neighbour before the putative induction window, recombine the recipient with that neighbour, and compare unrelated biological tissue. Rescue by the correct neighbour implicates an interaction. A permeable barrier or conditioned medium can distinguish a diffusible signal from a requirement for direct contact, but failure of conditioned medium is inconclusive if the signal is unstable, matrix-bound or requires a local concentration. A positive result through a barrier must also exclude passage of intact biological cells. Label sender and receiver separately so recipient conversion is not confused with expansion of contaminating donor biological cells.

Build a causal sender–signal–receiver chain. Delete a candidate ligand specifically in the sender, delete its receptor specifically in the recipient, and monitor an immediate pathway reporter before the terminal fate readout. Ligand addition should rescue ligand loss when the receptor is present; it should fail after verified receptor loss. Activated downstream machinery in the recipient can test bypass. Demonstrate a dose range that changes fate without changing survival, and use pulse-and-washout experiments to map a competence window. Loss of an endpoint after late treatment may indicate maintenance or differentiation failure rather than the original specification event. Independent genetic and pharmacological interventions reduce the risk of an off-target explanation.

Perform reciprocal heterotopic cell transplantation at several stages. Move a labelled prospective A biological cell into a B environment and a B biological cell into an A environment, with same-position grafts as surgical controls. Verify integration, survival and exposure to the host signal. If an early donor adopts the host fate, its potential is conditional at that stage. If a later donor retains its original fate while the host remains permissive and the graft survives, that supports cell determination. Retention in a host that never contacts or signals to the graft is uninformative. Follow mature morphology and function as well as markers; a hybrid marker state can represent incomplete conversion rather than either stable fate.

Regulative capacity needs another test. Ablate a mapped precursor and trace the source of replacement biological cells. Normal final anatomy is stronger evidence of regulative development if a previously distinct neighbour demonstrably adopts the missing fate; it is weaker if a few unablated members of the original lineage simply proliferate. Include sham ablation and damage-matched ablation elsewhere, quantify biological cell death and measure final biological cell numbers. Conversely, persistent loss of a structure after ablation does not prove mosaic development: the embryo may have lost an essential signal, physical scaffold or sufficient biological cell mass rather than an irreplaceable determinant.

A third mechanism, syncytial specification, must be tested if early divisions produce many cell nuclei in shared cytoplasm before cellularization. Image membranes to establish whether true cellular boundaries exist. Map nuclear position, local morphogen activity and nascent transcription; move cell nuclei relative to the gradient or shift the signal source while retaining nuclear identity. If equivalent cell nuclei adopt different programmes according to position in shared cytoplasm, that supports positional specification before biological cell boundaries. Perturbing cellularization alone is a poor test because it can damage several later processes.

An explicit interpretation matrix prevents overclaiming: retention after isolation supports specification; switching after a viable heterotopic graft shows continuing plasticity; retention after a verified conflicting graft supports determination; fate following a manipulated inherited determinant supports autonomy; and fate following a manipulated neighbour signal supports induction. These modes can coexist. An asymmetric maternal determinant might set competence for an external signal, and an induced fate can later become autonomous through positive feedback in a gene regulatory network. The final product should be a map of mechanisms and commitment times for particular lineages, not a single label for the entire organism.

A predictively useful simulation of a simple multicellular organism is a plausible long-term goal, but “fully” must specify resolution, environmental conditions and the predictions to be tested. A model that reproduces selected behaviors is attainable in a different sense from an exact simulation of every molecule throughout development, physiology, reproduction and aging. The former can be evaluated scientifically; the latter is neither established by present models nor a necessary definition of success.

A useful target might be a specified developmental stage of Caenorhabditis elegans, in a controlled environment, with predictions for cell states, body movement and responses to defined perturbations. Its small size, optical accessibility and extensively mapped cell lineage make it attractive. The classic embryonic lineage study traced divisions, cell deaths and fates, while also using perturbations to investigate the role of interactions.^[ref] Such a map constrains what a simulator must reproduce, but a list of normal outcomes does not itself specify the rules that generate them or how those rules respond to an unfamiliar intervention.

whole-cell modeling demonstrates both progress and the scale of the gap. A model of the bacterium Mycoplasma genitalium integrated 28 process models to connect cellular molecular functions with phenotypes, including growth, division and responses to genetic perturbation.^[ref] This is a substantial integration achievement, but it is a unicellular system with many experimentally supplied quantities and modeling assumptions. Its existence does not show that adding a collection of copies yields a validated multicellular organism. Differentiation, tissue organization, intercellular communication and mechanical interactions introduce new dependencies.

At the organismal scale, OpenWorm is a concrete attempt to integrate components of a virtual Caenorhabditis elegans, including neural activity, muscle activation, body mechanics and interaction with an environment. The project's published account describes a program of integration and validation, not a claim to have completed all the animal's biology.^[ref] A model can produce plausible locomotion while lacking developmental regulation, feeding or reproduction. Moreover, a neural wiring diagram alone leaves synaptic strengths, channel dynamics, neuromodulation, sensory transduction and the mechanical consequences of muscle forces underdetermined. Closing the loop through body and environment is therefore essential for behavioral prediction.

Several experimental developments make the goal more tractable. light-sheet fluorescence microscopy can follow living development with lower illumination of out-of-focus regions than conventional volume scanning, enabling repeated observation of cell divisions and movements, although phototoxicity and tracking errors still matter. single-cell RNA sequencing reveals heterogeneous cell states at scale; the lineage-resolved molecular atlas of Caenorhabditis elegans embryos connected transcriptional variation with known developmental ancestry.^[ref] spatial transcriptomics complements dissociated-cell measurements by retaining spatial context. These approaches constrain different aspects of a model rather than supply interchangeable readouts.

Measurement remains incomplete. A transcript count is not an absolute protein concentration, an enzyme activity or a reaction rate. Dissociation-based measurements usually destroy the sampled organism, so a trajectory reconstructed across many cells is not direct observation of every molecular state in one living animal. Mechanical properties, metabolite fluxes and signaling dynamics require other assays. Perturbations such as controlled changes in gene activity, laser ablation or altered environmental conditions are particularly useful because they distinguish causal mechanisms that fit the same unperturbed trajectory. A sufficient validation dataset must therefore contain interventions as well as descriptions.

Computationally, multiscale modeling offers a more plausible strategy than resolving every atom. A simulator might couple gene regulatory networks inside cells to cell growth, fate transitions and movement, then to extracellular transport and tissue mechanics. For a diffusing signal, for example,while each cell updates its internal state using the local signal concentration and its contacts. Those contacts change as the tissue moves, so the regulatory and mechanical descriptions must interact. Some processes can be deterministic at high copy number, while rare molecular events may require stochastic treatment. The choice of approximation should follow the prediction and timescale, rather than one mathematical formalism being imposed everywhere.

Platforms such as PhysiCell show how agent-based modeling can couple cell-level rules for division, death and mechanics to diffusible substrates and environmental conditions. Its demonstrations establish the feasibility of useful multicellular simulations, while leaving the user responsible for biological mechanisms and calibration.^[ref] Data-driven approaches add another tool: scGen demonstrated prediction of some held-out cellular perturbation responses. That supports a restricted kind of generalization, not unlimited prediction of arbitrary interventions, long-term development or an entire organism.^[ref] Learned components can be useful surrogates, but their training coverage and compatibility with physical constraints remain relevant.

The hardest obstacles are not just processor speed. Many parameter sets can reproduce the same measurements, creating identifiability problems. Errors in initial conditions, unobserved cell states and environmental boundary conditions can propagate through coupled modules. A mechanical model can change cell contacts, thereby altering signaling, which in turn changes growth and the mechanical state. Such feedback means that validating each module in isolation is necessary but insufficient. Interfaces must preserve units, spatial meaning, material balance and appropriate timing, and their numerical coupling must be verified. A larger model can accumulate more opportunities for error without producing more reliable predictions.

A credible development program would define success before fitting. For example, predict division timing, cell positions and a movement response within stated tolerances under a panel of withheld perturbations. Calibrate on different data, compare against simpler baselines, and evaluate whether predicted uncertainty covers observed variation across animals. Use sensitivity analysis and uncertainty quantification to identify which measurements would most improve the forecast, then design interventions that discriminate between competing mechanisms. model verification establishes that the simulator follows its rules; model validation tests those rules in the claimed biological domain. Neither a convincing animation nor recovery of data used in fitting is a sufficient substitute.

My expectation is therefore incremental predictive completion at chosen resolutions: increasingly broad, experimentally constrained models of development and behavior in controlled conditions. No fundamental physical separation prevents this, but limited observability, irreducible biological variation and the cost of resolving irrelevant detail prevent a guarantee of exact individual futures. The realistic goal is a whole-organism model with a declared and expanding domain of validated predictions, not an assertion that every detail has been captured because the simulation looks alive.

Quantitative lineage tracing shifted the interpretation of squamous renewal from a necessarily rigid stem-cell hierarchy toward population-level balance, while live-cell imaging and injury experiments revealed local coordination and context-dependent plasticity. The resulting picture is neither that every basal cell is identical nor that one universal hierarchy applies to every anatomical site. Methods measure different properties, and their assumptions help explain apparently conflicting conclusions.

A classical model of stratified epithelium places long-lived stem cells above a finite sequence of transit-amplifying cells, whose descendants leave the basal compartment and differentiate. Spatial columns of epidermal cells and slow label dilution were interpreted in terms of epidermal proliferative units. Such observations are compatible with a hierarchy, but do not establish it uniquely. A label-retaining cell may divide slowly, be terminally differentiated, or retain label because of assay kinetics. Conversely, an actively cycling stem cell can lose a dilution-sensitive label rapidly. Clonogenicity after culture or transplantation measures potential in a perturbed environment rather than necessarily identifying the cells that sustain undamaged tissue.

Sparse, inducible permanent clonal lineage tracing allowed the descendants of individual basal cells to be followed without relying on label retention. Clone numbers, basal-cell contents and differentiated descendants could be quantified at multiple chase times and compared with explicit stochastic models. In mouse epidermis, such analysis supported a model in which one progenitor class generated both renewing and differentiating daughters, with balance achieved across the population rather than one obligatory asymmetric division for every stem cell. ^[ref] Clones disappeared while surviving clones expanded, even though the tissue remained stable. That pattern is not evidence that disappearing clones were necessarily a separate short-lived class.

The bookkeeping is informative. Suppose a proliferative cell divides to give two proliferative daughters with probability $r$, one proliferative and one differentiating daughter with probability $1-2r$, and two differentiating daughters with probability $r$, where $0\le r\le 1/2$. The expected number of proliferative daughters is:Thus the proliferative population can maintain itself even when individual divisions have different outcomes. In a simplified critical branching process with division rate $\lambda$, the effective rates of clone-size increase and decrease are both $\lambda r$ per proliferative cell. Starting from one such cell:The product remains one: increasing mean size among surviving clones can accompany unchanged expected labelled cell number across all initial clones. These formulae illustrate survivor conditioning; actual tissue geometry, noncycling basal cells, finite compartments and local interactions can require a different quantitative model.

Work on mouse oesophageal epithelium combined label-retention assays with lineage tracing and found that a single proliferative population could account for maintenance. After wounding, its outcomes became biased toward producing proliferating daughters, then returned toward balance after repair. ^[ref] Changing the division rate and changing the probability of daughter fates are different responses. More divisions with unchanged balance can accelerate replacement without expanding the renewing pool; biased outcomes can expand that pool. Neither observation alone requires conversion of a terminally differentiated cell into a stem cell.

Evidence for heterogeneity also remained important. Marker-selected tracing and proliferation analysis in mouse epidermis supported distinct slow-cycling stem and committed progenitor compartments. ^[ref] Such results should not be dismissed simply because a single-progenitor model fits another dataset. Anatomical site, promoter expression, induction dose, chase duration and which cells are scored can change the sampled population. A driver that preferentially labels a slow-cycling subset and a broadly induced driver need not produce the same clone distribution. Models should be compared against joint measurements of clone survival, cell cycle kinetics and labelled differentiated output rather than one convenient summary statistic.

Live intravital imaging added information that destructive sampling cannot supply. Tracking the same cells shows which event precedes which: division, loss of basal contact, movement into a suprabasal layer, or neighbour expansion. In adult mouse epidermis, local differentiation was followed by neighbour expansion and compensatory cell cycle progression, supporting a mechanism in which cell loss locally stimulates self-renewal. ^[ref] Balanced stochastic clone statistics therefore need not mean that cells make independent random decisions insulated from their neighbours. A spatially coordinated system can generate population-level fate balance and apparent randomness at the level of individual retrospective clones.

Modern molecular methods refine, but do not replace, those functional tests. single-cell RNA sequencing identifies heterogeneous molecular states and pathway activity; spatial measurements locate them relative to basement membrane, follicles or wounds. A trajectory inferred from transcriptomes is not a directly observed lineage, and an intermediate expression profile can reflect a transient response rather than a cell reversing fate. Combining molecular profiling with permanent labels, repeated imaging and appropriately timed perturbations distinguishes state similarity, ancestry and causal requirement. cell cycle state and injury-associated inflammation also need to be separated from a genuinely distinct renewing population.

Direct evidence for cell fate plasticity requires more than changing proliferation within an existing progenitor class. Pre-injury labelling of GATA6-expressing differentiated sebaceous-duct cells showed that their descendants could move into wounded interfollicular epidermis, regain basal renewing properties and produce a broader range of epidermal descendants. Long-term tracing and live observations supported dedifferentiation under the wound conditions. ^[ref] The key comparison is the fate of the same pre-labelled population in undamaged and damaged tissue. It is not sufficient to find Gata6-lineage fluorescence in a cell that has just begun expressing the marker after injury.

Several controls make such a claim persuasive. Verify the initial labelled cell identities before damage; allow the inducer to clear; exclude rare pre-labelled stem cells or reporter leakage; and demonstrate sustained renewal rather than transient basal residence. Imaging a labelled cell's transition and showing its later functional output is stronger than a static double-positive image. Exclude cell fusion and selective expansion of an unrecognized initial subpopulation where relevant. Perturbing a proposed reprogramming regulator should preserve source-cell viability, since losing the source is not the same as preventing its identity change.

The general conclusion is a flexible hierarchy of evidence. Label retention measures division history, clonal tracing measures descendant output, live-cell imaging measures event order and local interactions, and molecular profiling measures state. Together they support dynamic population self-renewal, site-dependent heterogeneity and injury-induced epithelial plasticity, while leaving normal maintenance distinct from regenerative potential. Claims about one mouse skin region or oesophagus should be tested independently before being generalized to every squamous tissue or to humans.

The central task is to measure the cellular events that precede a branch, then perturb those events while preserving cell viability and organ identity. Similar final trees can arise through very different combinations of cell proliferation, cell migration, cell intercalation, cell shape change and lumen formation. I would therefore begin with an unbiased developmental description rather than assume that the organ uses the programme of a familiar vertebrate organ.

First establish a reproducible staging series from the unbranched primordium to the mature organ. Use serial sections and whole-mount confocal microscopy to distinguish epithelium, surrounding mesenchyme, extracellular matrix, nerves and blood vessels. Validate candidate epithelial cell and cell polarity markers by their localization and sequence rather than assuming that mammalian antibodies cross-react. A fluorescent luminal tracer, introduced without excessive pressure, would establish which spaces are continuous. Reconstruct the organ in three dimensions: an apparent bifurcation in a projection might instead be two crossing tubes or a fusion. Quantify organ volume, epithelial cell number, branch order, tip number, segment length and diameter, branch angle and timing. Distinguish lateral budding, tip bifurcation and branch fusion. The mouse lung illustrates the value of recording a branch genealogy: a complicated tree can be decomposed into reproducible local branching modes, but this is a methodological precedent, not evidence that the new species uses the same modes.^[ref]

Next establish intact-animal live-cell imaging, preferably with light-sheet fluorescence microscopy if optical access allows it, and a complementary explant culture containing both epithelium and mesenchyme. Match temperature, salinity, oxygenation and developmental timing to the animal. Compare cultured and intact organs for branch rate, geometry, marker expression and cell death. Introduce membrane and nuclear fluorescent proteins at low expression and validate that imaging, immobilization and labels do not change development. Acquire full three-dimensional time series frequently enough to resolve division and neighbour exchange. Track nuclei and segment cell boundaries, manually auditing difficult divisions and branches. Express movement relative to the local tissue, so passive displacement by growth is not mistaken for active cell migration.

Measure separately the sources of tissue expansion. Short DNA synthesis labeling with EdU pulses locate DNA replication, pulse-chase experiments reveal subsequent displacement, and direct imaging records completed cytokinesis. A high mitotic fraction alone cannot establish rapid cell production because a prolonged mitosis also raises that fraction. Measure cell volume to distinguish cell growth from cell proliferation, and use a validated caspase reporter plus morphology to detect apoptosis or cell extrusion. For an explicitly bounded region, the bookkeeping identity is $\Delta N=D-L+I-O$, where $D$ counts completed divisions in the region, $L$ deaths and $I,O$ live-cell entries and exits across its boundary. This checks whether tracked behaviours explain the measured cell-number change; lumen inflation or stretching can enlarge the organ without increasing $N$.

Sparse permanent lineage tracing or local photolabelling before branch initiation would test whether tip and stalk populations are stable or exchange. Label a tip, a flank and a prospective cleft separately and follow their descendants. Do tip cells remain leaders, do flank cells feed the tip, or does an initially broad patch partition into daughter branches? Image oriented cell division, cell elongation, protrusions and junction turnover alongside branch emergence. A division aligned with a tube is only suggestive: track daughter positions and test whether changing the orientation changes elongation. Quantify neighbour exchanges to identify cell intercalation. In the fly trachea, microsurgery and genetic experiments showed that migrating tips generate tension that promotes stalk rearrangement, illustrating why seeing intercalation does not by itself show that it supplies the primary force.^[ref]

Then make acute, staged causal tests. Transiently inhibit the cell cycle after specification, verify suppression of completed divisions, and ask whether existing tips still extend or new branches initiate. Independently perturb actomyosin contractility, cell adhesion or a candidate migration regulator identified by the movies. Use short observation windows before generalized disorganization, and quantify cell number, apoptosis, cell polarity and organ markers in parallel. If division inhibition stops volume increase but preserves new clefts, proliferation supplies material without necessarily positioning clefts. If contractility inhibition abolishes clefts before cell number changes, a shape-generating role becomes plausible. Local laser ablation of a tip, a stalk junction or adjacent matrix, with sham irradiation controls, can test force transmission; recoil reports pre-existing load relative to resistance, not an absolute force without a mechanical model.

Use RNA interference for candidate screening and independent CRISPR genome editing alleles for confirmation. Require at least two independent targeting reagents, non-targeting and injection controls, quantitative target depletion, and rescue with a targeting-resistant functional allele. Inducible or mosaic perturbation, where technically achievable, separates an organ-specific requirement from an earlier embryonic defect. Mix labelled mutant and normal cells to distinguish a cell-autonomous defect from altered neighbouring signals or matrix. Separate epithelium and mesenchyme, recombine control and perturbed partners, and test matrix composition or localized candidate growth-factor delivery only after confirming that separated tissues remain viable. Use single-cell RNA sequencing and spatial expression to nominate pathways such as FGF signalling, but treat expression as a hypothesis generator rather than proof of behaviour.

Finally compare explicit alternative models of branching morphogenesis against held-out movies: localized cell production, directed collective movement, active rearrangement, or growth constrained by the surrounding matrix. Each should predict both branch geometry and the observed cell trajectories, and its response to a spatially restricted perturbation. Randomize treatment, blind segmentation and use animals from independent clutches as biological replicates; many cells within one animal do not replace independent animals. The strongest outcome is a causal chain from molecular perturbation to altered cell behaviour to altered branch geometry, supported by rescue and by measurements showing that general developmental delay cannot explain it.

Epithelial cell fate plasticity is a context-dependent change in developmental potential or differentiated identity. It includes dedifferentiation toward a progenitor cell state, transdifferentiation between differentiated identities, and reversible intermediate states. It is distinct from selecting a pre-existing rare stem cell, and from ordinary progression down a lineage. epithelial-mesenchymal plasticity is one important form, but changing epithelial subtype without losing epithelial organization is also cell fate plasticity.

The strongest physiological rationale is that a tissue needs both stable specialization and reserve routes for replacement. During ordinary adult tissue homeostasis, a stem-cell hierarchy can efficiently supply predictable cell types. After injury removes a component of that hierarchy, local signals and newly available stem-cell niche space may permit a committed descendant to reacquire self-renewal. In mouse intestine, Alpi marked enterocyte-lineage progenitor cells normally generate short-lived absorptive descendants. After targeted ablation of Lgr5 positive intestinal stem cells, labelled progenitor cells can generate long-lived crypt-villus lineages with renewed stem-cell output. Comparing the same labelled population before and after ablation distinguishes injury-dependent cell fate plasticity from its normal fate. The experiment concerns progenitor cells in the enterocyte lineage; it should not be generalized to every terminally mature villus enterocyte.^[ref]

A complementary model is the mouse airway. Genetic depletion of basal stem cells, combined with prior lineage tracing of luminal secretory cells, revealed secretory-cell descendants that acquired basal-cell properties and sustained subsequent epithelial production. Experiments with isolated cells and cocultures further showed that contact with basal cells inhibits this conversion. cell fate plasticity is therefore regulated by the tissue environment, not simply an intrinsic capacity unmasked whenever a marker changes. Its interpretation requires controls for residual basal cells and for whether the secretory-lineage driver labelled basal cells before injury.^[ref]

Development provides additional contexts. Embryonic epithelia can undergo epithelial-to-mesenchymal transition during processes such as gastrulation, and mesenchymal-to-epithelial transition can subsequently assemble epithelial structures. These are regulated changes in adhesion, cell polarity, motility and gene expression, rather than a simple binary switch read from E-cadherin alone. In wound repair, transient relaxation of epithelial organization can facilitate migration while retaining intercellular contacts. A partially migratory repair state need not entail a complete mesenchymal conversion or a permanent change of lineage. Following wound closure, restoration of barrier function and polarized architecture is as important as the initial ability to move.

The most convincing experimental design labels a defined starting population irreversibly before the inducing event, allows the inducer to clear, then follows labelled descendants through the transition. Check immediate labelling specificity, include no-inducer and uninjured controls, quantify labelling efficiency and rule out ectopic driver activation after injury. Demonstrate the new state with multiple markers, morphology and function, and persistence after the inducing signal is removed. Serial epithelial production or a second injury challenge tests stable self-renewal; a transient appearance of a basal marker does not. Sparse labelling and continuous live-cell imaging help distinguish conversion from contamination, cell fusion or selective expansion of a rare initially unrecognized progenitor cell.

single-cell RNA sequencing can identify intermediate cell states and candidate regulators, but inferred trajectories are not lineage histories. Combine it with genetic barcodes, spatial localization and timed perturbation of candidate transcription factors or niche signals. Organoids offer defined media, prospective cell isolation and manipulation of matrix stiffness. A sorted population's ability to make a structure in culture establishes potential under those conditions; dissociation and a permissive medium may induce a capacity rarely used in vivo. Compare culture results with intact-tissue fate mapping before claiming a normal physiological contribution.

The evolutionary argument is a trade-off, rather than a claim that every plastic response is an adaptation. Conditional access to alternative cell fates could buffer environmental injury and permit repair without maintaining a separate reserve stem-cell pool for every possible lesion. Conserved signalling modules can also be redeployed in new developmental contexts. These advantages must be weighed against loss of specialized function, energetic costs and the danger of inappropriate persistent growth. Comparative experiments should ask whether homologous lineages undergo similar transitions under comparable injuries; shared expression of a few regulators is insufficient to prove that an ancestral cell fate plasticity programme has been conserved. At the tissue level, cell fate plasticity supplies alternative states on which selection can act, but it does not itself establish heritable evolutionary change.

Clinically, controlled cell fate plasticity may improve epithelial repair, provide transplantable cells and make patient-derived disease models. The same capacity can undermine treatment: a tumour may switch to a less drug-sensitive lineage or regenerate a stem-like population after therapy. In carcinoma, epithelial-mesenchymal plasticity can contribute to invasion and treatment resistance, but neither a single EMT marker nor its absence establishes the mechanism of metastasis. In a lineage-traced mouse lung-cancer model, the tested EMT programme was dispensable for lung metastasis while contributing to chemoresistance; the conclusion is specific to that model and reporter coverage.^[ref] The useful clinical aim is therefore a defined, reversible transition that restores function, with evidence that it does not leave a persistently proliferative or malignant population.

A learned method for predicting certain cellular perturbation responses from single-cell RNA sequencing data. Generalization to evaluated held-out cell types or conditions does not establish predictive validity for all interventions.^[ref]