<%@ page contentType="text/html; charset=iso-8859-1" language="java" import="java.sql.*" import="java.io.*" %> iViS FAQs <% java.util.Date now = new java.util.Date(); %>

iViS (In Vivo - In Silico) - Grand Challenge Website

UEA | JIC | IFR | CMP
Grant Challenge | Proposal | State of the Art | In Vivo - In Silico | Expected Outcomes | Work Plan
Ronan Sleep | Andrew Bangham | Enrico Coen
UEA Projects | JIC Projects | IFR Projects
Projects | Laboratories | People | Papers | Books | Journals | Online Content
Biological | 3D Modelling | Cell Modelling | 4D Data Structures
Notepad | Reports | Literature Review | Software Development | Data Stuctures | 3D Modelling
Email List | Contacts
Email List

iViS FAQs

In Vivo In Silico - Frequently Asked Questions
  1. What is the iViS Challenge?
  2. Sounds like artificial life - what's different?
  3. Why might I be interested in iViS?
  4. How did iViS start and what's its status?
  5. What is an iViS roadmap?
  6. What are some current iViS roadmap themes?
  7. What sort of people and places are doing iViS things?
  8. Is there an iViS masterplan?
  9. Is there an iViS pot of gold?
  10. Is it driven by scientific curiousity, or big bucks?
  11. What achievements will indicate iViS success?
  12. Does it promise a revolutionary shift in the accepted paradigm of thinking or practice?
  13. Won't the iViS vision happen anyway because of work already underway in the life sciences?
  14. Will its promotion as a grand challenge contribute to the progress of science?
  15. Does it have the enthusiastic support of established scientific communities?
  16. Does it appeal to the imagination of the general public?
  17. What kind of long-term benefits to science industry, or society may be expected?
  18. Does it have international scope?
  19. How does the project split into sub-tasks or sub-phases, with identifiable goals and criteria, say at five-year intervals?
  20. What calls does it make, for collaboration of research teams with diverse skills?
  21. How can it be promoted by competition between teams with diverse approaches?
  22. Why has it been so difficult so far?
  23. Why is it now expected to be feasible in a ten to fifteen year timescale?
  24. What are the first steps?
  25. What are the most likely reasons for failure?

What is the iViS Challenge? iViS aims to realise fully detailed, accurate and predictive models of some of the most studied life forms used as models in biology. Examples include Aribidopsis, bakers yeast (S. Cerivisiae) and the Nematode worm (C. elegans). This approach builds on partial computer models that are already under development many laboratories. The aim is to create a complete, consistent, integrated representation of all that is known about a particular plant or animal. The result should be a computer embodiment of all that is known about the life form. It should be accessible to humans via extensible view selection mechanisms that include the interaction modes possible between an experimenter and the real life form, and also between the life forms themselves. Thus for an animal model, such as C. elegans (the Nematode worm) it should be possible to view phenomena such as:

  • DEVELOPMENT from an initial fertilized cell to a full adult, at various resolution levels. An accurate model will respect knowledge about, for example: cell lineage, cell differentiation, cell lifetime, morphology, size and relation between major cellular sub-systems. Virtual experiments (e.g. moving a virtual cell during development, or making an incision) should lead to the same outcomes as real life.
  • CELL FUNCTION and INTERACTION: the specific functions of cells should be captured in appropriate detail together with principal modes of interaction.
  • MOTILITY and SENSORY aspects of behaviour: types of reaction to various stimuli, including neighbouring life forms; speed and nature of movement.
  • ENVIRONMENTAL INTERACTION: interactions between organisms and the surrounding environment should be captured.

Sounds like computer animation and virtual life - what's different? The big difference is that the iViS animations and models have to correspond precisely to the way nature works. The designers of computer animations and virtual life forms can get away with something that looks good in a limited context. An iViS model has to be able to stand up to probing by the world's best life scientists. And eventually, we hope iViS models may tell the life scientists things they don't know which are later confirmed by experiment.

How did iViS start, and what is its status? iViS is a guiding vision rather than a specific project. It started as one of a number of Grand Challenge proposals designed to stimulate UK Computer Science research. At present, iViS is no more than a website and some ideas in a few people's heads.

Why take an interest in iViS ? Some reasons might be:

  • To get in on the race to construct the first convincing in-silico model of a whole life form.
  • To see what's going on in the field of in-silico modelling.
  • To see if some of your ideas might be useful to others.
  • To see if you can solve some of the problems other people are having.
  • To get together with others to make coherent research proposals driving towards a recognised goal.

What is an iViS roadmap? An iViS roadmap is a particular interpretation of the iViS vision, a focussing theme, and a list of milestones with approximate timescales by which progress can be judged.

Some iViS roadmap themes Some examples are:

  • Developmental Biology / Morphogenesis: focus on modelling the processes of growth of a life form from initial cell to adult.
  • Single Cell: concentrate on a single cell lifeform.
  • Biochemical process architecture: concentrate on the mechanical details of life, viewing it as an interaction between three abstract virtual machines: the gene machine: says what protiens to make when and where, and directs membrane construction and protein embedding; the protein machine makes the proteins, which signal conditions and events to the gene machine and implements fission and fusion; and the membrane machine which hosts reactions, holds receptors, actuators and genome, and confines regulators.

Is there an iViS masterplan? iViS itself has no masterplan, but individuals working towards its goals may have their own masterplans. The FAQs about roadmaps give some idea of the sort of plans people are working on.

Who does iViS things, and where? Follow the people and links URLs..

Is it driven by scientific curiousity, or big bucks? Scientific curiosity is a big motivation. The relationship between machines and living things is an ancient but ongoing preoccupation of man, crystallised by von Neumann, Turing, Wiener and others. There may be big bucks in it too, but as with the web it's not easy to predict where..

What achievements will indicate iViS success? Some specific signs of progress are indicated below:

  • Showpiece demonstrators launched for selected life forms.
  • Widening use of iViS modelling frameworks as a medium of exchange between life scientists in their everyday work, leading to a de-facto move from using traditional data base technology to hold knowledge about life forms to the use of hi fidelity iViS models instead.
  • Conversely, the Computer Scientists will take inspiration from Biology to construct new ways of specifying complex reactive systems that construct and maintain themselves from small initial and perhaps sketchy specifications: we just need to understand how nature manages to engineer the worm, the weed and the bug from a single cell without detailing all the intermediate construction stages.

Does it promise a revolutionary shift in the accepted paradigm of thinking or practice? It might well produce revolutions in thinking both in computing and systems sciences and also in the life science. Just as structured programming brought order and discipline to the management of massive software programs, so the use of advanced computational models and languages for process interaction can bring order and discipline to structuring the life science data mountain. Conversely, Computer Scientists can hope to develop theories and techniques for the analysis and synthesis of massively complex reactive systems. Nature clearly supports adaptive parasitic assembly processes (e.g. the growth of a plant from a seed using environmental energy and other resources), as evidenced by the development of any plant or animal from a simple seed. Understanding how this works may have profound repercussions for industries ranging from systems engineering to manufacturing.

Won't iViS happen anyway because of work already underway in the life sciences? No, because most projects are concerned with only part of the picture. The role of iViS is to identify opportunities where a small amount of additional effort could develop a whole organism model from existing work. Noone is likely to make money out of the creation of iViS, but everyone is likely to benefit. Just like the web..

Will thepromotion of iViS as a grand challenge contribute to the progress of science?
Yes. For computing scientists, life form modelling provides a challenging but clear and concrete target. It complements the many theoretical and pragmatic 'free' explorations underway of topics such as ubiquitous computing, artificial life, emergent properties of complex systems. More speculatively, success may produce a quite revolutionary new development-as-computing paradigm. For life scientists, the introduction of advanced system languages and architectures, together with appropriate notions of abstraction, may help tame the bewildering complexity found in nature.

Does it have the enthusiastic support of established scientific communities? The use of model lifeforms dominates the life sciences. For example, there are over 300 laboratories working on C. elegans, so there is considerable interest in modelling it. There is a similar level of activity on Aribidopsis and S. Cerevisiae. A number of groups are already active in computer modelling, for example:

Does it appeal to the imagination of the general public?
A complete simulation of a living creature appeals to the public interest in the creation of artificial life. Already computer simulations of cancer growth are being used to help decide treatment regimes. As the accuracy of modelling increases, so models will become a key part of the doctor's armoury in the fight against disease.

What kind of long-term benefits to science industry, or society may be expected?
For Science: significant acceleration of progress due to:

  • Massively improved shared distributed computational observation engines based on standard models for recording, representing and accessing knowledge about life forms
  • Effective frameworks for incorporating knowledge about complex systems with applications in both biology and systems engineering
  • Inspiration to create radically new models of computation
For Society: just as accurate computer models of hydrodynamics have almost entirely replaced live nuclear testing, so we might hope that sufficient investment in accurate biological models might remove or at least considerably weaken many of the arguments for experimenting with live animals. Bringing together our rapidly expanding knowledge about the mechanics of cell lifetime control in the form of a predictive model may give us an indispensable tool for sharing and creating new knowledge about diseases.

Does it have international scope? A large number of labs are beginning to build serious models of parts of plants or animal systems, but some coordination is needed for trustable whole models to emerge. iViS can play a leading part in this.

How does the project split into sub-tasks or sub-phases, with identifiable goals and criteria, say at five-year intervals? iViS will proceed by idenfitying a number of direction of work for integrative biology. The major criterion for inclusion as an iViS strand is that the route aims at whole life form modelling. Obviously a whole iViS strand will consist of substrands working on separate aspects, but there must be a clear coordinating mechanism. Some - probably rather optimistic - milestones of progress of a typical iViS strand are given below:

  • 5 years : partial but accurate development models of exemplar organisms beginning to predict the result of experiments.
    10 years : complete models for some aspects beginning to emerge
    15 years : first complete models

What calls does it make, for collaboration of research teams with diverse skills? Successful completion of the challenge will need efficient distributed interactive simulation coupled with a good visual / haptic front end extended to model the desired experimental environments. Specifying and realising accurate models even with just hundreds of cells will stretch the current technology, and achieving the necessary realism at the interface is beyond the current state of the art. We can expect continued advances in the general capability for distributed simulation and graphics to emerge from industry, driven for example by the games market. The real challenge will be in determining the right design choices for representing cells and cell interactions. There are several groups already working in this difficult area, and it looks as if there is enough evidence on which to base some initial design decisions for the challenge. However, to achieve the best chance of success will require an inspired mix of knowledge about life science and computing.

How can it be promoted by competition between teams with diverse approaches? Good science is about good questions. A critical step towards a successful challenge is to identify sharp sub-questions that different teams can attack simultaneously, perhaps in the form of a competition. A good recent example of the use of a competition to advance science is the Abbadingo One DFA (Deterministic Finite Automata) Learning Competition5. This involved careful attention to establishing a proving ground consistent with but just beyond the state of the art, and precise rules for submission and victory. Some preliminary competition regarding the architecture for the whole design might be attempted, and this might also help draw in people with the necessary vision and other qualities to drive the challenge.

Why is it now expected to be feasible in a ten to fifteen year timescale? Some ad-hoc computer modelling techniques are already making headway by concentrating on specific aspects. The real challenge is to do a small number of whole organisms very, very well, and to synthesise a generic framework capable of dealing with a wide range of knowledge about a wide range of creatures.

What are the first steps?

  1. Identify stakeholders, building on existing BioModelling communities (e.g. MIPNETS) with EPSRC/BBSRC assistance.
  2. Establish a series of meetings to flesh out the skeletal challenge into a much more precise form, and construct a roadmap for the challenge.
  3. Preparation of more detailed outline plan and presentation to major stakeholders.
It is suggested that the initial focus is on modelling accurately the development of a small number of life forms. This focus is proposed because:
  • There is considerable detail about some well-studied models such as C. elegans, arabidopsis, yeast etc.
  • The number of cell divisions modelled gives a logarithmic measure of the computational cost, so enabling the modelling work to concentrate on establishing detail for some number of divisions which is easily within reach of fairly standard computers.
  • Getting even the first division right requires establishing the key attributes of cells and their interactions.
  • Last but not least, the initial divisions of many classes of life show similarities.
  • The exact life forms chosen should emerge from detailed consideration of the issues to be tackled and the work already underway. For example, work is already underway on C. elgans.

What are the most likely reasons for failure? probably cultural rather than technical::

  • Vested interests are likely to cause tensions within and between existing scientific cultures, even (especially?) within the same discipline. Strong and visionary efforts will be needed to establish an effective framework and to attract the stronger rather than the weaker scientists to a risky interface area.
  • Biologists may need to adapt to new approaches to representing knowledge that cuts across existing biological compartments, and to take on board some sophisticated notions from distributed computing.
  • Computer scientists may have to suppress a natural desire to see the challenge as a way of justifying yet more work on some existing research topic that may be of only peripheral relevance to the challenge.
  • We continually underestimate the rate of progress in computing, and this frequently leads to investment in research efforts that address problems, which, by the end of the research, have largely vanished. A technology intercept strategy is needed which works towards a model which can be realised with technology which will be around in 5,10,15 years time.
©2004 Ioannis Elpidis / Ronan Sleep