**[Full text of the proposal](https://www.synapse.org/Portal/filehandle?ownerId=syn5659209&ownerType=ENTITY&xsrfToken=1EA1466FCA55F7EAE33833333900F1BC&fileName=Idea5.pdf&preview=false&wikiId=414654)** ###Anonymous Review 1 and Authors Response _ **Impact: ** Being able to discover the signals on the surface of cells that are behind morphogenesis would significantly improve our understanding of multicellular structures. How individual cells are aware of the overall structure, and how they then regulate gene expression to produce the appropriate cell type and local structure is still poorly understood. A result here would have impact in developmental biology, as well as practical applications such as agriculture._ _**Feasibility: ** Intercellular signally is a complex and poorly understood domain, and is further complicated by environmental factors. The proposed model seems to be too simple to be able to explain the signalling for the broad range of phenomenon the authors hope to explore._ **Response:** This is correct. But we think that a modeling which pretends to find the most general important law(s) of development should start from the simplest (but not trivial) idea, and next be continued towards higher levels of complexity. The idea of the existence of a generic epigenetic code on a cell surface, if proved experimentally to be correct, can be a basis for both mathematical formalization allowing to find this law and a search for biological pathways implementing this code into a proper geometry via the proper chain of cell events.   _**Overall evaluation:** On many potential experimental systems, use fluorescent probes to detect the spatial distribution of glycoconjugate residues on a cell surface, and to track changes over time. It is unclear from the proposal how the seemingly disparate experimental systems were chosen, and why they are all expected to be somehow amenable to fitting by the proposed model._ **Response:** These experimental systems were chosen in order to detect possible changes in the distribution of glycoconjugate residues in a course of each particular type of ?cell event?. Namely, 1) Xenopus embryogenesis, stating from the zygote- for the cell event _Division_. 2) Somatic embryogenesis of dycotelodone plants- mostly for the cell event _Division_ at the early stages, and for the _combination of_ cell events _Division_, _Growth_ and _Differentiation_ at the later ones. A culture of plant somatic embryos has the advantage that it can be synchronized and thus have a lot of embryos at the same stage of development. Also, a culture of plant somatic embryos is a good biological model for the comparison of the behavior of glyco-markers in the case of normal and abnormal development, which in a certain percentage of cases usually happens in the culture. 3) A process of assini formation from stem cells in breast tissues- mostly for the cell event _Apoptosis_, which is an essential part of this process, and also for checking the specificity of the code distribution during an asymmetrical division of stem cells. 4) A process of myoblast differentiation in muscles ? for the cell event _Differentiation_.   _It seems that the proposed coarse grained model of only 8 quadrants for spatial distribution of residues would not be able to capture the intricate structures of multicellular organisms. The images in the supplement seem to suggest more resolution than this, which may be because the geometric model for cell structures could have more detail than the model for surface signals. One potential resolution would be to consider a model system that is geometrically more constrained in terms of structure, such as only growing on a 2D plane, or a regular grid._ **Response:** Even with 8 octants (which is the simplest case with which we start our proof-of-concept) we can capture the complexity of structures of multicellular organisms. Having octants with vectors of 8 (or more) types of coding molecules in each, and having a function on a set or (subsets of them) which determines cell events, allows for generating complex geometries in three dimensions.   _Another confusion is regarding the conjecture that surface residues are the cause of the multicellular structure._ **Response:** "conjecture that surface residues are the cause of the multicellular structure" ? is not exactly what we meant. The surface residues give a code for a cell according to which this cell will undergo a "cell event", with or without taking into consideration a set of signals from the cell surrounding, which are also a function of the surface codes of the corresponding surrounding cells. If a surface code gives a cell (e.g., a zygote) an instruction to divide, and then gives it to the cell descendants ? then it will be a multicellular structure; if not ­? we will have a unicellular organism.   _Perhaps the authors can clarify the literature on the different types of intercellular signalling, and why surface residues are particularly relevant to the the proposed study structures._ **Response:** The set of statements supporting the idea that the cell surface information (glyco- residues, or any other possible type) may be relevant for the proper **coordination** of all intra- and inter-cellular cell events for developing a proper complex pattern, is presented in the Introduction of our proposal. We do not pretend for now to suggest molecular mechanisms underlying this coordination, "intercellular signaling" being one of them. We are trying to suggest a model, describing the idea and to find the general rules of the surface code behavior by mathematical formalization. We use modeling in order to better explore this theoretical conjecture and, in a way, as a proof-of-concept.

Created by Chloé-Agathe Azencott caz
###Anonymous Review 3 and Authors Response _**Impact: ** As far as I can judge the proposal is ambitious and aims at a fundamental question. _ _**Feasibility: ** See comments below re evaluation. _ _**Overall evaluation:** The proposal addresses a fascinating problem. It proposes a set of experiments (detecting glycoconjugate residues on cell surface) driven by theory and simulations. The modelling approach has interesting methodological component based on hidden-state models for spatial distribution._ _Two main concerns about components of the proposal that should be thought out more deeply._ _(1) It is not clear how the parameters of the models will be determined. Will they be learned/estimated from data? (how?) or estimated based on prior/other experiments? or searched for and then compared to collected data?_ **Response:** The preliminarily idea (which is in real work always subject to reevaluation and to change) is that the predictions on the parameters (e.g. the boundary conditions) will be given by a theoretical search in the frame of the mathematical part. Afterwards, during the sets of suggested experiments, the real values of parameters will be found inside the predicted intervals.   _(2) Experimental validation. The results of Figure 3 are impressive. I am concerned about the lack of clear quantitative methodology to compare simulations results with the collected data._ **Response:** Correct. We are working on it.   _ One issue for example is that in this setup, due to the large number of analyzed cells, the statistical problem of multiple comparisons is challenging, it may be easy to find patterns that "look similar" but are not discriminative enough about the model._ **Response:** In our theoretical mathematical part which was not presented in the proposal due to the limited size, we pay a lot of attention to this problem. Namely, we suggest to use a metric for the comparisons of two objects (organisms) by using Gromov-Hausdorff type distance.  
###Anonymous Review 2 and Authors Response _ **Impact: ** If successful, the project would lead to better understanding of cell development and differentiation process, and potentially to the elucidation of new "laws" governing morphogenesis._ _**Feasibility: ** The study aims at collecting data that could help validate the mathematical model proposed by the authors. Collecting data in itself seems feasible. Designing specific experiments to test the model, such as specifying particular distributions of glycoconjugate residues on cell surfaces, looks more challenging._ _**Overall evaluation:** The project is interesting, but I found the preliminary results a bit weak to justify that the mathematical model worked out by the authors is likely to work._ **Response:** Our project consists of 3 parts ? a theoretical one (a set of ideas converted into mathematical formalization), modeling one and an experimental one (for which we are trying to find support). The modeling part, which we have included in the proposal, is based on the very simplified version of the theoretical part. We use it for better exploring our theoretical conjectures and as a possible proof-of-concept. We are planning to move towards a more complex and sophisticated model driven by the development of the theoretical part together with necessary experimental data.  

Idea 5: Molecular code for geometry of morphogenesis page is loading…